2014-10-01 06:05:21 morning 2014-10-01 06:05:36 sisaacs: buildrepo 2014-10-01 06:05:52 will build the entire world 2014-10-01 06:06:27 blueness: we gave up pth 2014-10-01 06:06:34 its not portable (despite of its name) 2014-10-01 06:06:56 npth works though 2014-10-01 09:12:43 typo.. whoops 2014-10-01 09:34:03 ncopa, monit v5.9 is out 2014-10-01 09:35:25 vkrishn: updated. thanks 2014-10-01 09:35:43 I thing rsync on buildlogs wold be useful for remote quick view 2014-10-01 09:38:35 as a.o has rsync , a folder logs/bld[12] can be added, binding on available nfs 2014-10-01 11:00:18 ncopa, can npth be used with gnupg-2 as a drop in? I'm not familiar with npth 2014-10-01 11:01:45 blueness: nope 2014-10-01 11:01:57 gnupg-2.0 is a mess 2014-10-01 11:02:12 gnupg-2.1 should work with npth though 2014-10-01 11:02:26 ncopa, that's good enough thanks! 2014-10-01 11:02:36 what happened with alpine was 2014-10-01 11:02:46 we added gnupg 1.x 2014-10-01 11:02:52 to be able to move to musl 2014-10-01 11:03:04 then we added gnupg-2.1 which is beta 2014-10-01 11:03:18 pth and gnupg-2.0 was just not worth it 2014-10-01 11:05:22 ncopa, i got pth to build, but i'm not sure it was working 2014-10-01 11:05:44 because of my hackiness so i'm more comfortable with gnupg-2.1 + npth 2014-10-01 11:07:39 i find it ironic that pth is "GNU portable threads" 2014-10-01 11:07:51 and happen to be extremely non-portable :) 2014-10-01 11:08:01 well 2014-10-01 11:08:12 the idea was that you port only pth to get app portable 2014-10-01 11:08:27 but the sad part is pth itself was not so easily portable, and unmaintainable 2014-10-01 11:08:29 a portability layer 2014-10-01 11:08:49 and in many ways lame at what it tried to accomplish 2014-10-01 11:09:27 ncopa, did we get the usbkbd/usbmouse kernel config change in 3.0-stable ? 2014-10-01 11:09:37 not yet 2014-10-01 11:10:00 do we get it before next 3.0.z release? 2014-10-01 11:10:16 i think its the motivating factor for 3.0.5 release 2014-10-01 11:10:34 once kernel is updated in 3.0-stable we can do 3.0.5 i think 2014-10-01 11:10:45 ok. good. i'd like to get that sometime soonish. no panic, but within a week or two. 2014-10-01 11:10:50 there was some other changes 2014-10-01 11:10:55 specifically the x86 2014-10-01 11:11:01 and the randomize base 2014-10-01 11:11:13 right 2014-10-01 11:11:14 that i'm a bit sceptic to 2014-10-01 11:11:25 but, not enabling it causes significantly more work 2014-10-01 11:11:45 so i think I'll end up just cherrypick things from edge 2014-10-01 11:11:51 as they are 2014-10-01 11:13:12 ncopa, yes, the code that was failing was implementation dependant. i forget now the details 2014-10-01 11:13:24 ucontext likely 2014-10-01 11:13:53 blueness did you get firefox running for x86? 2014-10-01 11:14:35 ncopa, no i haven't been wowkring much on gui stuff 2014-10-01 11:14:57 k 2014-10-01 11:15:05 i have some issues with it 2014-10-01 11:15:06 i have a partial desktop going based on xfce4, but i was too busy to push through enough stuff to make it a decent desktop 2014-10-01 11:15:12 oh i can believe it 2014-10-01 11:15:25 it works ok with x86_64 2014-10-01 11:15:39 oh, other thing that we havent yet fully solved is nfs 2014-10-01 11:16:10 they do some nasty assumptions about getaddrinfo implementation 2014-10-01 11:16:20 which i think need refactor 2014-10-01 11:16:45 ncopa, where's the problem, in musl or nfs? 2014-10-01 11:16:50 nfs 2014-10-01 11:16:55 nfs-utils 2014-10-01 11:17:29 and they do assumtions on some /proc file system 2014-10-01 11:17:39 fabled sent a patch for that upstream 2014-10-01 11:17:59 it's combination of all three 2014-10-01 11:18:05 musl, nfs-utils and kernel 2014-10-01 11:18:19 apparently my patch is not correct as it breaks v3; but made v2 work with musl 2014-10-01 11:18:27 nfs v3 that is 2014-10-01 11:18:33 yeah 2014-10-01 11:18:40 shoud clean up the patch and split it parts 2014-10-01 11:18:59 we also need enable dnotify in our kernel (which i dont want) or add inotify support to nfs-utils 2014-10-01 11:19:04 shouldnt be too hard 2014-10-01 11:19:21 ncopa, what's the issue with dnotify? 2014-10-01 11:19:39 just that i dont have it enabled 2014-10-01 11:19:45 the API is not that good 2014-10-01 11:19:49 uses signals 2014-10-01 11:19:52 has problems 2014-10-01 11:19:58 so most apps dont use it 2014-10-01 11:20:27 only nfs-util's idmap hasnt ported to inotify 2014-10-01 11:20:37 there was a thread on it on the ml this spring 2014-10-01 11:21:32 it looked relatively simple, at least the concept 2014-10-01 11:21:56 dnotify uses signals, inotify uses a file handle 2014-10-01 11:22:04 but they use libevent 2014-10-01 11:22:30 so it should just be to create an fd for inotify and plug it to libevent 2014-10-01 11:23:03 i've only used inotify 2014-10-01 11:24:52 but i must say 2014-10-01 11:25:04 working with musl have been *very* pleasant 2014-10-01 11:25:06 on multiple levels 2014-10-01 11:25:40 code is clean and nice 2014-10-01 11:26:03 the are very strict to standards 2014-10-01 11:26:20 ncopa, did your nfs-utils patches get accepted, or are still pending? 2014-10-01 11:26:20 they prefer do what is technically correct 2014-10-01 11:26:52 fabled: they got all accepted, except one - the getaddrinfo thingy 2014-10-01 11:27:00 but that was bad and need rework 2014-10-01 11:27:02 awesome 2014-10-01 11:27:18 yup 2014-10-01 11:27:34 i should upstream our rpcbind and libtirpc patches 2014-10-01 11:28:57 -rc4 is tagged with those too 2014-10-01 11:29:12 oh, 1.3.1 is released 2014-10-01 11:29:20 it is? 2014-10-01 11:29:26 yeah 2014-10-01 11:29:29 at least tag is in git 2014-10-01 11:30:02 5 days ago 2014-10-01 11:30:39 i should rebase my patch, and split it 2014-10-01 11:30:42 try to fix issues it has 2014-10-01 15:26:33 Has anyone attempted ZFS on AlpineLinux? 2014-10-01 15:31:48 How about replacing mdev with udev? (ZFS depends upon udev) 2014-10-01 15:39:56 ZFS works fine I guess, and the patches made by the guy porting are being upstreamed 2014-10-01 15:40:08 you can use some ancient udev as well 2014-10-01 15:46:47 sisaacs: i dont think zfs depends on udev 2014-10-01 19:45:49 barthalion: Sorry I missed your earlier comment. Do you know if I can contact the guy porting ZFS? 2014-10-01 19:54:54 sisaacs: I guess he posted his patches on mailing list 2014-10-01 19:55:20 and if not, some of his changes are in aports tree, so there must be his mail somewhere 2014-10-01 20:13:07 i guess you're looking for me 2014-10-01 21:05:47 <[xming]> I think I use too much bash :p 2014-10-02 04:11:19 great news! KDE is working on Alpine :) 2014-10-02 04:56:12 HlArvEe5nyd5: ah, I haven't noticed you on nicklist, sorry 2014-10-02 04:56:15 sisaacs: ↑ 2014-10-02 06:04:11 k0r10n: cool! 2014-10-02 06:04:15 congrats! 2014-10-02 08:36:08 ncopa: about chromium, almost done with thit (successfully compiled 12400/15293 files :D), but i'm stuck on breakpad 2014-10-02 08:39:14 maybe could you help with it, because i don't have such experience with C :/ 2014-10-02 08:41:27 i think we hacked breakpad on firefox 2014-10-02 08:42:49 ok will look at it, guys from google redefined all syscall macros... 2014-10-02 08:44:53 there are no any patch in firefox aport 2014-10-02 08:45:00 *patches 2014-10-02 08:45:21 except duckduckgo 2014-10-02 08:45:28 its in xulrunner 2014-10-02 08:45:39 ok, ty 2014-10-02 08:51:17 barthalion: sorry myself. i change nicks regularly, but they're always 12 char random strings. 2014-10-02 09:32:40 <[xming]> moin, more bash bump http://sprunge.us/DWiK 2014-10-02 10:11:08 <[xming]> there are missing imgs on http://wiki.alpinelinux.org/wiki/Main_Page 2014-10-02 11:08:15 [xming]: missing imgs? 2014-10-02 11:10:33 <[xming]> on the main page there should be 6 big images 2014-10-02 11:10:54 <[xming]> I only see image for developer documentation 2014-10-02 11:11:16 <[xming]> e.g. Installation is missing http://wiki.alpinelinux.org/wiki/1.5x 2014-10-02 11:11:45 <[xming]> img src == a href 2014-10-02 11:12:06 <[xming]> hmm nvm 2014-10-02 11:12:14 <[xming]> looks fine on FF 2014-10-02 11:13:04 they have been missing for me long time 2014-10-02 11:13:12 but i think somebody mention they work. 2014-10-02 11:13:15 <[xming]> it's all messed up in my chromium 2014-10-02 11:13:48 <[xming]> works 100% in FF 2014-10-02 11:14:10 <[xming]> ncopa: more bash bump http://sprunge.us/DWiK 2014-10-02 11:14:13 yes you are right. ff is ok 2014-10-02 11:15:02 hm 2014-10-02 11:15:12 i thought 027 fixed those... 2014-10-02 11:15:14 :-( 2014-10-02 11:15:26 http://bugs.alpinelinux.org/issues/3407 2014-10-02 11:15:33 <[xming]> I suspect that more bash patches are coming 2014-10-02 11:15:42 bash is top quality 2014-10-02 11:15:50 <[xming]> all the fixes are some how patchy 2014-10-02 11:16:13 <[xming]> to bash or not to bash? 2014-10-02 11:16:48 well, writing a shell with the complexity of bash is not easy 2014-10-02 11:16:51 I read this article http://www.itnews.com.au/News/396256,further-flaws-render-shellshock-patch-ineffective.aspx 2014-10-02 11:17:09 however... when i was doing work on GNU as an intern in highschool, i can tell you the people they had working on it 2014-10-02 11:17:09 <[xming]> I don't think that 27 fixed CVE-2014-7186,CVE-2014-7187 2014-10-02 11:17:14 didn't know what they were doing 2014-10-02 11:17:22 which seems to indicate that CVE-2014-7186,CVE-2014-7187 was fixed in 027 2014-10-02 11:17:25 and there's probably a ton of amateur mistakes in there relating to strings 2014-10-02 11:17:42 if i were to guess, i would say this bash thing is going to play out for months 2014-10-02 11:17:50 kaniini: i belive you are right about that 2014-10-02 11:18:14 thanks to it got so much media attention, it will be get a critical review 2014-10-02 11:18:20 <[xming]> only months? 2014-10-02 11:18:27 that is an optimistic view 2014-10-02 11:18:34 and there will be found lots of ugliness 2014-10-02 11:18:37 <[xming]> IMHO idd 2014-10-02 11:18:47 if you want my more mopey estimates, 18 - 24 months with the eventual discontinuation of bash as a product 2014-10-02 11:19:09 <[xming]> I love bash, but just a shell and for my own scripts, not for spawning other things 2014-10-02 11:19:20 i doubt it will go away 2014-10-02 11:19:39 it will already be a harmful thing for GNU's reputation (which is already in a bad place in terms of robustness and code quality) 2014-10-02 11:20:19 <[xming]> the moment they moved from v2 to v3, it became so bloated that I had to eliminate bash as much as possible for /bin/sh 2014-10-02 11:20:39 i like fish 2014-10-02 11:20:46 it is a kind of heavy C++ app though 2014-10-02 11:21:00 <[xming]> saw that years ago, it was funny :D 2014-10-02 11:21:01 but as far as a user-shell, it is nice 2014-10-02 11:21:15 problem is that bash is used for scripting 2014-10-02 11:21:22 has support for arrays etc 2014-10-02 11:21:37 <[xming]> idd, I just wrote a ini parser yesterday in bash :/ 2014-10-02 11:22:00 <[xming]> even hash tables 2014-10-02 11:22:53 <[xming]> with bash4 it's possible to write script in pure bash in most cases without awk/sed/cut/tr/.... 2014-10-02 11:23:18 welp 2014-10-02 11:23:43 the more complexity you add to an already bad base 2014-10-02 11:23:54 means you're widening the chance of getting 0day'd significantly 2014-10-02 11:24:42 <[xming]> idd, but it's so much fun 2014-10-02 11:25:11 <[xming]> and as user shell, it's okay until there is a local priv escalation 2014-10-02 11:25:24 anyway, i would expect a 'security-oriented' bash fork coming soon 2014-10-02 11:25:42 <[xming]> I hope so 2014-10-02 11:26:23 <[xming]> bash used to be so cool in the early 90s 2014-10-02 11:27:17 <[xming]> chromium really mis render the wiki 2014-10-02 11:59:33 [xming]: i figured that bash is ~700-800kb 2014-10-02 11:59:37 lua is ~300kb 2014-10-02 12:00:16 so lua is faster and smaller and has cooler tables/arrays 2014-10-02 12:01:07 <[xming]> but to interact with the system I still find a shell easier 2014-10-02 12:01:35 that is true 2014-10-02 12:02:11 <[xming]> doing many files cp/mv/ln/... shell is much easier than any real programming language 2014-10-02 12:02:12 copy/move files etc is easier with a shell 2014-10-02 12:02:16 yes 2014-10-02 12:02:34 <[xming]> that what I needed and also a config file 2014-10-02 12:02:46 <[xming]> so I wrote a mini ini parser in pure bash 2014-10-02 12:02:56 :) 2014-10-02 12:03:18 <[xming]> writing a config file parser is easier in any language than shell 2014-10-02 12:04:00 yup 2014-10-02 12:04:17 <[xming]> fitting things in a limited one dimensional hash/array was not very obvious 2014-10-02 12:08:47 <[xming]> ncopa: I like golang better than lua, but it's huge 2014-10-02 12:49:32 hi@all 2014-10-02 12:50:00 did anyone got issues with svn with ipv6 on alpine 2.7? 2014-10-02 12:50:03 hi. which apk dependencies does one need to compile the current 3.16.3 stable kernel from source? alpine-sdk does NOT suffice. 2014-10-02 12:50:37 StarWarsFan: i have a vauge memory about issues with ipv6 and subversion 2014-10-02 12:50:49 iirc ipv6 was used and it didnt fall back to ipv4 2014-10-02 12:51:14 no probs with "localmodconfig" on ubuntu/mint, on alpine it's a nightmare with missing "bc" or whatnot. help PLEASE! drives me nutz 2014-10-02 12:51:32 the following works as expected on 3.0 but not on 2.7: 2014-10-02 12:51:33 fou: depends="mkinitfs linux-firmware" 2014-10-02 12:51:33 makedepends="perl sed installkernel bash gmp-dev bc" 2014-10-02 12:51:34 svn co https://ssl.bulix.org/svn/lcd4linux/trunk 2014-10-02 12:51:43 nocpa: got both 2014-10-02 12:51:44 on 2.7 i got this: 2014-10-02 12:52:13 fou: do you have perl sed installkernel bash gmp-dev bc 2014-10-02 12:52:20 ~/tmp svn co https://ssl.bulix.org/svn/lcd4linux/trunk 2014-10-02 12:52:20 svn: E000101: Unable to connect to a repository at URL 'https://ssl.bulix.org/svn/lcd4linux/trunk' 2014-10-02 12:52:20 svn: E000101: Error running context: Network is unreachable 2014-10-02 12:52:34 but that's wrong 2014-10-02 12:52:38 i think gmp-dev was needed for grsecurity gcc plugin 2014-10-02 12:53:07 because if i try to make wget, it looks like this: 2014-10-02 12:53:23 ~/tmp wget https://ssl.bulix.org/svn/lcd4linux/trunk 2014-10-02 12:53:23 --2014-10-02 14:52:48-- https://ssl.bulix.org/svn/lcd4linux/trunk 2014-10-02 12:53:23 Resolving ssl.bulix.org... 2001:41d0:8:995c::1, 5.39.87.92 2014-10-02 12:53:23 Connecting to ssl.bulix.org|2001:41d0:8:995c::1|:443... failed: Network is unreachable. 2014-10-02 12:53:23 Connecting to ssl.bulix.org|5.39.87.92|:443... connected. 2014-10-02 12:53:24 HTTP request sent, awaiting response... 301 Moved Permanently 2014-10-02 12:53:27 ... 2014-10-02 12:53:41 so it seems that svn fails on the first try using ipv6 2014-10-02 12:54:23 StarWarsFan: yes, i think I have seen that too 2014-10-02 12:54:36 i dont remember what the proper fix is/was 2014-10-02 12:54:49 it might be i just rmmod ipv6 2014-10-02 12:55:11 it sounds like uclibc is misbehaving 2014-10-02 12:56:26 hm 2014-10-02 12:56:38 nocpa: i'll reboot and check every single dependcy you gave me. why does documentation does SUCK SO MUCH on this site when everything else on your distro just plain rockz. 2014-10-02 12:57:08 thats actually a good and valid question :) 2014-10-02 12:57:20 fou: what error do you get? 2014-10-02 12:57:22 one that crossed my mind a dozen times today 2014-10-02 12:57:31 the exact error message you get 2014-10-02 12:58:17 reason the documentation for compiling 3.16 sucks is because youre the first (that i know) to do it 2014-10-02 12:58:24 they probably introduced some new dependency 2014-10-02 12:58:37 make localmodconfig complains "crc32c config something not". have to reboot to reproduce. thanks for help. hopefully notz brb 2014-10-02 12:58:53 3.16 compiles flawless on mint17. no issues. no nothing no introduction 2014-10-02 12:59:02 some base dependency just lacking in alpine 2014-10-02 13:04:20 root@builder07:~ rmmod ipv6 2014-10-02 13:04:20 rmmod: can't unload 'ipv6': Resource temporarily unavailable 2014-10-02 13:04:45 but lsmod shows it as used by the bridge!? 2014-10-02 13:07:33 if someone achieved compiling the 3.16.3 vanilla "stable" kernel WITH a minimal alpine installation, I'd be very delighted to learn how and with what apks. thank you so much 2014-10-02 13:08:15 make localmodconfig produces "crc32c config not found!!" and subsequently anything else does not work either. 2014-10-02 13:10:45 some not horrendously outdated "kernel compiling on alpine howto" that does not point to already decaying gentoo-docu would be great too! :) 2014-10-02 13:12:49 fou, 'abuild deps' should install all dependencies you need 2014-10-02 13:13:32 fou can you get APKBUILD from kernel-vanila and abuild build ? 2014-10-02 13:14:17 http://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package 2014-10-02 13:14:27 crow: i just want to download the stable kernel and build it on a minimal alpine install without setting up bloated development environments on ubuntu or such or "CREATING PACKAGES" 2014-10-02 13:14:42 when make bzImage does not suffice Im not interested in it 2014-10-02 13:17:15 duno.. you can check apkbuild file to see what is done in package and try to reproduce.. http://git.alpinelinux.org/cgit/aports/tree/main/linux-vanilla/APKBUILD?id=8b4939898592ffa22fb2ea436aa73258ada75fad 2014-10-02 13:18:03 here you have config files http://git.alpinelinux.org/cgit/aports/tree/main/linux-vanilla?id=8b4939898592ffa22fb2ea436aa73258ada75fad 2014-10-02 13:18:11 your're nuts 2014-10-02 13:18:43 mate i get my config from /boot or /proc/config.gz 2014-10-02 13:18:53 then i clean that shit up with olddefconfig and localmodconfig 2014-10-02 13:19:05 i can do all that on shell with just make 2014-10-02 13:21:26 alpine-sdk didnt even include bc which is obviously needed for "make localmodconfig". so what apks else do i need to just "make bzImage" please? thank YOU! 2014-10-02 13:23:48 fabled replied you already 2014-10-02 13:24:09 can help you with that, i would try in one vm to build an alpinelinux kernel and see what packaged are pulled and if it compiles fine, then would try make bzImage... 2014-10-02 13:24:15 cant* 2014-10-02 14:03:01 ncopa, refreshed nfs-utils patches 2014-10-02 14:03:19 there was one bug fixed that _might_ have caused the nfsv3 issues 2014-10-02 14:03:39 so it might be idea to give it a test spin if anything improved (or got worse) 2014-10-02 14:03:46 i sent them to linux-nfs list too 2014-10-02 14:04:31 ok 2014-10-02 14:04:35 good 2014-10-02 14:49:07 ncopa, http://wiki.alpinelinux.org/wiki/Alpine_Linux:IRC_Availability 2014-10-02 14:49:49 still wip, I would be adding some 'if and buts' as no devs are bound to it, they are just guidelines 2014-10-02 14:50:51 once done would add a link to http://wiki.alpinelinux.org/wiki/Alpine_Linux:IRC 2014-10-02 14:51:34 nice 2014-10-02 14:51:35 thanks 2014-10-02 20:05:16 how can i remove the ipv6 module? 2014-10-02 20:05:29 on a.l v2.7? 2014-10-02 20:56:06 StarWarsFan: remove? 2014-10-03 05:23:48 morning 2014-10-03 05:23:53 clandmeter: right 2014-10-03 05:25:00 see backlog here starting at "[14:50:00] did anyone got issues with svn with ipv6 on alpine 2.7?" 2014-10-03 05:25:39 ncopa told me "[14:54:49] it might be i just rmmod ipv6" 2014-10-03 06:45:57 StarWarsFan: you can blacklist it and reboot? 2014-10-03 06:58:43 StarWarsFan i did not check if this can be done in alpinelinux. but on archlinux i am using this: GRUB_CMDLINE_LINUX_DEFAULT="quiet ipv6.disable=1" inside /etc/default/grub , i dont have access to alpinelinux box right now to check if such file does exist. 2014-10-03 07:28:15 ncopa: looks like openrc doesnt support recent kernels? 2014-10-03 07:28:22 i see a message about cgroups 2014-10-03 07:28:27 ok? 2014-10-03 07:28:48 kernel has policy to not break userspace 2014-10-03 07:28:55 so it should just work 2014-10-03 07:29:10 i get an error on boot 2014-10-03 07:29:27 runscript.sh 2014-10-03 07:29:33 line 215 2014-10-03 07:30:27 what error do you get? 2014-10-03 07:31:13 ncopa: https://dl.dropboxusercontent.com/u/18871200/2014-10-03_093011.png 2014-10-03 07:31:18 seems like you can set RC_DEBUG=yes in /etc/rc.conf 2014-10-03 07:31:36 this is first boot after setup-alpine 2014-10-03 07:31:47 vanilla iso 3.0.4 2014-10-03 07:31:47 v3.0? 2014-10-03 07:33:01 x86 or x86_64? 2014-10-03 07:33:05 64 2014-10-03 08:39:23 ncopa: ? 2014-10-03 08:40:56 clandmeter: i tried to start the vanilla vm 2014-10-03 08:41:08 but virt-manager didt display checkboxes 2014-10-03 08:41:10 gtk3 broke themes 2014-10-03 08:41:27 so i ended up fixing the adwaita gtk theme... 2014-10-03 08:41:32 :) 2014-10-03 08:41:38 np 2014-10-03 08:41:53 i was just wondering where you went. 2014-10-03 08:46:05 clandmeter: the cgroups error does not show up when booting alpine-vanilla-3.0.4 iso 2014-10-03 08:46:42 gtk3 is so great that audacious is switching to qt 2014-10-03 08:46:44 ;) 2014-10-03 08:49:30 ncopa: well i didnt do anything, just boot it. 2014-10-03 08:49:35 maybe its diff on vmware 2014-10-03 08:50:23 kaniini: yeah, many does that 2014-10-03 08:52:32 xsetroot: command not found 2014-10-03 08:52:46 do we have that in aports? 2014-10-03 08:52:55 probably 2014-10-03 08:53:32 ubuntu has it in x11-xserver-utils 2014-10-03 08:55:26 oh we have it 2014-10-03 08:55:33 we have 2014-10-03 08:55:39 its a single pkg 2014-10-03 08:55:59 and i also checked up debian/ubuntu packages :) 2014-10-03 08:56:13 only to find that apk search xsetroot finds it for me :) 2014-10-03 08:56:55 yeah, sometimes things are just to obvious :) 2014-10-03 08:58:06 i'm not able to reproduce the cgroups issue 2014-10-03 08:58:13 crow: thx but grub is not an option as there are some XEN oder LXC-containers 2014-10-03 08:58:21 clandmeter: what do you mean with "blacklist"? 2014-10-03 08:59:37 StarWarsFan: add it to /etc/modprobe.d/blacklist.conf 2014-10-03 09:00:22 ah ok, will try this 2014-10-03 09:00:23 thx 2014-10-03 09:00:30 ncopa: i can try it on real hardware and see if is repoducable 2014-10-03 09:13:57 http://weblog.rubyonrails.org/2014/9/12/Rails-4-1-6-and-4-0-10-has-been-released/ 2014-10-03 09:14:04 they say they drop support in dec 2014-10-03 09:14:22 we have rails 4.0 in our v3.0 branch 2014-10-03 09:14:45 we cannot give rails support after that, inc sec issues 2014-10-03 09:14:54 qemu is a similar situatin 2014-10-03 09:15:08 we cannot give support for it anymore 2014-10-03 09:16:08 i wonder if we should do like ubuntu and have a 'main' and 'universe' repo 2014-10-03 09:16:34 and we only provide support/sec fixes for 'main' 2014-10-03 09:19:29 we should drop redmine and rm -rf *ruby* 2014-10-03 09:20:22 agree 2014-10-03 09:20:23 /rant mode off 2014-10-03 09:20:30 https://lists.fedoraproject.org/pipermail/devel/2014-October/202876.html 2014-10-03 09:20:40 fedora considering dropping bash as /bin/sh 2014-10-03 09:20:48 i saw the headlines yes 2014-10-03 09:20:49 bash's slow death begins ;) 2014-10-03 09:21:20 clandmeter: hm, not working. still on the output of lsmod 2014-10-03 09:21:36 root@builder07:~ lsmod 2014-10-03 09:21:38 ... 2014-10-03 09:21:44 ipv6 305642 77 bridge 2014-10-03 09:22:42 ideally we should fix uclibc 2014-10-03 09:22:59 but feels bad to patch uclibc now that we have musl :-/ 2014-10-03 09:23:10 i'd rather fix nfs for musl 2014-10-03 09:23:53 a workaround would be fine for me as 3.0 is definately the future 2014-10-03 09:24:06 but we need to build 2.7 packages too, so svn is a must have 2014-10-03 09:24:22 i understand the problem 2014-10-03 09:24:45 sorry for interrupting. but what is the situation with qemu? 2014-10-03 09:24:47 so actually i got major issues on both systems, 2.7 and 3.'0 :-( 2014-10-03 09:31:28 StarWarsFan: http://subversion.1072662.n5.nabble.com/IPv6-URL-support-in-SVN-v1-8-4-works-with-IPv4-not-with-IPv6-td185549.html 2014-10-03 09:32:15 https://lists.fedoraproject.org/pipermail/devel/2014-October/202906.html 2014-10-03 09:32:16 haha 2014-10-03 09:32:48 HlArvEe5nyd5: the problem with qemu is that they dont provide any longterm support 2014-10-03 09:33:02 It is also funny to hear the systemd author talk about not breaking 2014-10-03 09:33:03 things and creating lots of work. 2014-10-03 09:33:05 there are sec issues with qemu-2.0.x 2014-10-03 09:33:05 ACTION falls over laughing 2014-10-03 09:33:17 but no fixes from upstream, as upstram only supports 2.1 2014-10-03 09:33:25 ncopa: i know this posting but my resolv.conf contains only one entry 2014-10-03 09:33:26 oh. 2014-10-03 09:33:40 and going to qemu-2.1 is no immediate option i guess. 2014-10-03 09:33:55 StarWarsFan: yes, but i think what happens is the uclibc resolver figures out that therea re ipv6 addresses 2014-10-03 09:34:17 HlArvEe5nyd5: i did upgreade qemu to 2.1 in alpine v3.0 2014-10-03 09:34:24 and i just hope it does not break anything.. 2014-10-03 09:34:43 its just not possible to provide any longterm support for it :-/ 2014-10-03 09:37:53 i see :/ 2014-10-03 09:38:27 ok 2014-10-03 09:38:28 enough autism 2014-10-03 09:38:32 ACTION quits reading fedora-devel 2014-10-03 09:39:33 observation: since bhyve's interface is similar to kvm's /dev/kvm interface, might it be possible to make the bhyve emulator part run on kvm? 2014-10-03 09:40:04 for well-behaved virtio guests, this may be interesting 2014-10-03 09:41:15 StarWarsFan: we could try figure out what is going on in apr 2014-10-03 09:41:51 anything you think to do... 2014-10-03 09:43:05 it is likely related how apr uses getaddrinfo or similar 2014-10-03 09:51:16 StarWarsFan: i think the quick-and-dirty solution woudl be to add ipv6.disable=1 to boot option or add ipv6 to /etc/modprobe.d/blacklist 2014-10-03 09:51:30 you probably need to reboot to get rid of ipv4 2014-10-03 09:51:33 ipv6 2014-10-03 09:51:48 tried this without success 2014-10-03 09:52:02 but this was a xen container 2014-10-03 09:52:08 will try on a real machine now... 2014-10-03 09:52:48 i can reproduce it in a container here 2014-10-03 09:53:16 the entry must look like this, right? 2014-10-03 09:53:18 blacklist ipv6 2014-10-03 09:53:23 yes 2014-10-03 09:53:25 not more, not less 2014-10-03 09:53:25 ok 2014-10-03 09:53:37 also check that ipv6 is not in /etc/modules 2014-10-03 09:55:23 there is only the entry "af_packet" 2014-10-03 09:57:11 ok, bare-metal it works :-) 2014-10-03 09:57:31 so only the xen container is left 2014-10-03 10:01:22 ha 2014-10-03 10:01:26 i found a workaround 2014-10-03 10:01:52 StarWarsFan: do ip addr 2014-10-03 10:02:06 you will see that you have some ipv6 auto assigned 2014-10-03 10:02:11 manually delete those 2014-10-03 10:02:27 ip -6 addr del ::1/128 dev lo 2014-10-03 10:02:47 ip -6 addr del fe80::fc79:ceff:fe62:77a1/64 dev eth0 2014-10-03 10:02:58 i did those 2 things and then it started to work 2014-10-03 10:03:17 interesting, will try right now... 2014-10-03 10:05:47 right, this works 2014-10-03 10:05:54 but is this persistent during reboot? 2014-10-03 10:12:44 probably not 2014-10-03 10:12:55 i guess its auto aadded when bringing up the interface 2014-10-03 10:13:20 but you can add it to rc.local 2014-10-03 11:03:59 StarWarsFan: maybe you could use the ipv4 ip addr instead of hostname 2014-10-03 11:07:53 ncopa: ? 2014-10-03 11:07:57 what do you mean by that? 2014-10-03 11:12:39 ;; ANSWER SECTION: 2014-10-03 11:12:39 https://ssl.bulix.org. 23 IN A 5.39.87.92 2014-10-03 11:13:05 svn co https://5.39.87.92/svn/lcd4linux/trunk 2014-10-03 11:25:33 hm, sounds like a bad idea to put an ip address into an APKBUILD 2014-10-03 11:26:12 breaks the build if it changes... 2014-10-03 11:59:34 fabled about your rpi image from here http://dev.alpinelinux.org/~tteras/ , how do you prepare an SD card with this? 2014-10-03 12:07:40 crow, create SD, format it FAT and extract the tarball there 2014-10-03 12:08:07 the tarballs are regular alpine run-from-tmpfs setup 2014-10-03 12:08:34 rpi works so that the bios has bootloader that understands FAT, so it reads named files from the SD 2014-10-03 12:08:44 no special bootloader preparation needed 2014-10-03 12:10:42 fabled ok great let me try it. 2014-10-03 12:10:54 erm i think i saw msg from you that you wanna update tarbals? 2014-10-03 12:11:21 " (15:21:01) (@fabled) should probably update rpi tarball at somepoint" :) 2014-10-03 12:16:32 yes 2014-10-03 12:16:36 oh right 2014-10-03 12:16:43 forgot to do that 2014-10-03 12:16:47 i'll do it now 2014-10-03 12:17:09 greate :) 2014-10-03 12:19:46 i'll just do a quick test boot of the image and upload it (it's built) 2014-10-03 12:22:10 thank you. 2014-10-03 12:24:56 uploaded http://dev.alpinelinux.org/~tteras/alpine-edge-141003-armhf.rpi.tar.gz 2014-10-03 12:29:46 i will try it later, my rpi is not here. i have time to play with. 2014-10-03 14:56:48 ncopa: do you think its possible to push openrc update (modloop change) to 3.x? 2014-10-03 15:01:12 is it well tested? 2014-10-03 15:01:47 i think it should be ok 2014-10-03 15:02:16 clandmeter: could you please file a bug with 3.0.5 as target with soem explanation why it needs to be fixed in v3.0 2014-10-03 15:02:21 and i'll look at it on monday 2014-10-03 15:02:45 ok, i hope i dont forget :) 2014-10-03 15:02:48 im going home 2014-10-03 15:02:52 me too 2014-10-03 15:02:56 have a nice weeked! 2014-10-03 15:02:57 nice weekend! 2014-10-03 20:20:34 <[xming]> This is hopefully the final bash patch for the known CVEs 2014-10-03 20:20:50 <[xming]> ncopa: http://sprunge.us/iQUL 2014-10-03 20:22:55 <[xming]> it seems that 27 just shielded the parser from untrusted input, 28 actually patches 2 CVEs and 29 patches the last known one? 2014-10-03 20:23:03 <[xming]> http://lcamtuf.blogspot.se/2014/10/bash-bug-how-we-finally-cracked.html 2014-10-03 20:23:07 <[xming]> just for the heads up 2014-10-03 20:25:23 yeah, basically 028 is Florian's patch 2014-10-03 20:25:37 lcamtuf's CVEs are unexploitable IIRC 2014-10-03 20:25:42 but 029 fixes one of them 2014-10-03 20:36:00 <[xming]> hehe bash2 received patches after 11 years :p 2014-10-03 20:57:22 n8@all 2014-10-04 04:16:19 hi, after upgrade to linux-grsec-3.14.19-r3 xen reboots in a loop after "Loading Dom0" - no way to get any error-msg. booting without xen works expected the erros that it could not load modules / conecting to xen. Is there any reported error known? btw. 3.14.17 was running perfect! 2014-10-04 11:57:11 <[xming]> http://sprunge.us/iQUL 2014-10-04 12:59:19 it's not final fix 2014-10-04 12:59:47 029 fixes CVE-2014-6277 2014-10-04 13:01:12 there is one more left 2014-10-04 13:10:07 Can a github project be packaged like this: http://sprunge.us/RcFj 2014-10-04 13:10:44 I wonder what I should use as pkgver 2014-10-04 13:12:51 The build deps aren't yet complete :-) 2014-10-04 13:19:53 <[xming]> then this http://sprunge.us/DeSA 2014-10-04 13:21:21 <[xming]> jomat: pkgrel is for the local version, when you do chances with the same upstream version you need to bump the pkgrel 2014-10-04 13:21:34 <[xming]> s/with/wthin 2014-10-04 13:25:33 [xming]: Yeah, sure, pkgrel=0, but i'm not sure what to use as pkgver :-) 2014-10-04 14:03:05 https://github.com/irungentoo/toxcore/blob/master/toxcore/crypto_core.c#L5 2014-10-04 14:03:07 oh dear 2014-10-04 14:03:24 https://github.com/irungentoo/toxcore/blob/master/toxcore/crypto_core.c#L39 2014-10-04 14:03:25 bad 2014-10-04 14:03:43 using special intrinsics like that allows for observations to be made, thus reopening the timing bug 2014-10-04 14:04:14 https://github.com/irungentoo/toxcore/blob/master/toxcore/crypto_core.c#L88 2014-10-04 14:04:24 no evidence of using secure memory management methods 2014-10-04 14:05:36 oooh, nice. i have other problems with tox: https://github.com/irungentoo/toxcore/blob/master/docs/updates/Crypto.md#crypto-request-packets 2014-10-04 14:05:47 sending pubkeys of the peers is great for traffic analysis 2014-10-04 14:06:20 i wouldn't use this 2014-10-04 14:06:24 it is probably some NSA joke 2014-10-04 14:07:31 i am using tox for some months as an example that even nacl/libsodium does not prevent you from writing crap 2014-10-04 14:08:03 i'm really confused here on this 2014-10-04 14:08:16 as far as i can see 2014-10-04 14:08:25 there's no protections against replay attacks with the comms 2014-10-04 14:08:35 and nonce is sequential, which should never be 2014-10-04 14:08:54 ooh. nonce is ok if sequential 2014-10-04 14:08:56 it's a nonce 2014-10-04 14:09:04 the only thing is to use it once. 2014-10-04 14:09:07 it's "ok" but it's a bad practice 2014-10-04 14:09:08 can be sequential 2014-10-04 14:09:17 no. totally ok. only reuse is bad. 2014-10-04 14:09:42 * NOTE: This code has to be perfect. We don't mess around with encryption. 2014-10-04 14:09:44 this is still funny 2014-10-04 14:09:52 indeed :) 2014-10-04 14:10:20 jomat: there is tox in testing already 2014-10-04 14:10:29 is there? 2014-10-04 14:10:40 i'll be sure to let our security officer know we're distributing NSA-ware 2014-10-04 14:10:51 yeah, some old snapshot I pushed months ago 2014-10-04 14:10:57 can you remove it 2014-10-04 14:11:12 sure, there is no use of it anyway 2014-10-04 14:11:17 this is not software that should be recommended for use 2014-10-04 14:11:22 it's way too old to be usable and I don't care about it 2014-10-04 14:12:08 Follows pretty much the principle of the torrent DHT: http://www.bittorrent.org/beps/bep_0005.html (READ IT) 2014-10-04 14:12:11 But: Vastly simplified packet format and encryption. 2014-10-04 14:12:12 oh dear 2014-10-04 14:12:34 after I wasted time trying to explain to ncurses client upstream why bumping version in Makefile is not enough to call something a release I lost interest 2014-10-04 14:12:35 i'm not sure how you 'simplify' kademlia 2014-10-04 14:12:47 kademlia is literally 2 ops 2014-10-04 14:12:51 search and ping 2014-10-04 14:13:13 there's tons of weasel-words in this documentation like that 2014-10-04 14:13:19 which is more evidence that it's NSA-ware 2014-10-04 14:13:47 -If there are bad clients and/or the list isn't full: 2014-10-04 14:13:49 -Send a ping request to the client 2014-10-04 14:13:51 yay, DoS vector 2014-10-04 14:14:07 this stuff is really impressive fail 2014-10-04 14:14:18 -Send nodes 2014-10-04 14:14:21 -If the ping_id matches what we sent previously (check in the corresponding pinged list.): 2014-10-04 14:14:23 -Each node in the response is pinged. 2014-10-04 14:14:31 yay, they completely turned kademlia into a DoS vector 2014-10-04 14:15:08 it's fucking kademlia 2014-10-04 14:15:11 i mean, ffs 2014-10-04 14:16:07 HlArvEe5nyd5: i hope no actual activist organization is using this software, it seems really bad at a cursory glance 2014-10-04 14:16:56 actually i have a list of "rules of thumb against cybersnakeoil" and if you apply that, you'll be left with less than a handfull tools that survive. 2014-10-04 14:17:20 which is too bad, because if it were competently done, it could be useful for activists as they wish to be 2014-10-04 14:17:40 post has been translated here, and contains the rules 1-6 (7th has been added after translation was done) https://rsdy.github.io/posts/security_and_the_rise_of_snakeoil.html 2014-10-04 14:17:41 but i suspect tox is actually a case where, the cure is more dangerous than not 2014-10-04 14:18:35 rule 7: neglects the sorry state of endpoint security. 2014-10-04 14:19:12 yes, i agree with those guidelines 2014-10-04 14:20:55 what i don't understand is 2014-10-04 14:21:02 i'm tempted also to add rule 8: has not matured against the multi-year scrutiny of the security industry. 2014-10-04 14:21:05 where in libtoxcore is the keypair generation ? 2014-10-04 14:21:16 their code is such a mess i've yet to find it 2014-10-04 14:21:44 mostly its done with nacl.crypto_randombytes 2014-10-04 14:21:52 ?! 2014-10-04 14:22:11 ok, without the nacl. prefix 2014-10-04 14:22:17 but, you have to do more than that when you're using EC systems 2014-10-04 14:22:26 that's how you do it with nacl/libsodium 2014-10-04 14:22:32 the privkey must describe a valid curve point 2014-10-04 14:22:39 yeah 2014-10-04 14:22:41 no 2014-10-04 14:22:42 if it doesn't, you've shot yourself in the foot 2014-10-04 14:22:53 the private key is 32byte integer 2014-10-04 14:23:09 wich you apply to the function crypto_scalarmult_base 2014-10-04 14:23:15 which gives you a point on the curve 2014-10-04 14:23:20 and is the public part 2014-10-04 14:23:25 the scalar is in fact your secret 2014-10-04 14:23:38 right, but 2014-10-04 14:23:49 ACTION is the maintainer of pysodium and a bunch of other nacl/sodium based tools.... 2014-10-04 14:24:00 at least with some curves, you have to do more than that 2014-10-04 14:24:08 i guess they are using djb's curve which is fine with whatever 2014-10-04 14:24:12 yeah. this is the beauty of djbs implementation. 2014-10-04 14:24:31 safecurves.cr.yp.to for more info and also nacl.cr.yp.to 2014-10-04 14:24:58 i even ported libsodium to arm cortex-m3-s :) 2014-10-04 14:26:03 either way, it's still interesting to find the key management code in libtoxcore 2014-10-04 14:26:11 to see if it practices even remotely decent key hygeine 2014-10-04 14:26:29 based on what i have seen thus far, i am sure it dumps some binary blobs with chmod 777 or something stupid 2014-10-04 14:27:03 pls do some review very interested in any findings. 2014-10-04 14:28:44 /* The main loop that needs to be run at least 20 times per second. */ 2014-10-04 14:28:48 oh dear 2014-10-04 14:29:05 well 2014-10-04 14:29:17 my conclusions thus far are 2014-10-04 14:29:22 - it's GPL3 2014-10-04 14:29:36 anyway i wish someone did a proper implementation of the pond protocol. some friends looked at the current go implementation and said it could be improved. 2014-10-04 14:29:44 which is bad for their goals, because they want to be accessible 2014-10-04 14:29:46 but the protocol is nice 2014-10-04 14:30:07 their core libs really shouldn't be GPL 2014-10-04 14:30:10 that's just bad 2014-10-04 14:32:52 HlArvEe5nyd5: yes, pond uses axolotl ratchet which is good 2014-10-04 14:32:57 actually i was able to compile pond on alpine with go. so maybe it could be packaged. 2014-10-04 14:33:41 actually just last week i implemented the axolotl ratchet based on libsodium: https://github.com/stef/saxolotl 2014-10-04 14:36:54 tox looks very messy, i said it 2 years ago when i first heard about it, and i'll say it now 2014-10-04 14:37:31 the guys behind it clearly don't have the necessary software engineering skills to make it a success 2014-10-04 14:38:03 oh, success does not depend on skills but on marketing and hitting the right chords with the demand side of the market.... 2014-10-04 14:38:09 look at apple... 2014-10-04 14:50:41 was reading musl docs, will adding --disable-static during build be better 2014-10-04 15:38:12 <[xming]> jomat: LIBTOXCORE_LT_VERSION=0:0:0 :p 2014-10-04 15:49:23 Oh dear, I didn't look under the hood yet /o\ 2014-10-04 15:55:30 But I'll abandon tox for now and have a look at https://github.com/ricochet-im/ricochet :-D 2014-10-04 16:01:12 :) 2014-10-04 16:01:22 so now we have to scrutinize ricochet? 2014-10-04 18:59:57 -da-:-) 2014-10-04 19:00:03 ups 2014-10-04 19:00:07 yeah :-) 2014-10-04 19:36:36 <[xming]> jomat: why not retroshare? 2014-10-04 19:39:30 jomat: or xmpp+otr? 2014-10-04 19:39:56 I'm not convinced to retroshare 2014-10-04 19:40:00 retroshare? have you checked that hodgepodge ever? 2014-10-04 19:41:04 rs doesn't look better than tox… 2014-10-04 19:42:04 actually it looks worse :) 2014-10-04 20:22:44 barthalion, [xming]: actually my girlfriend is a big fan of video telephony... As skype is a no-go (also for her), and jitsi is a pain, i hoped tox would be a good alternative 2014-10-04 20:30:48 meh 2014-10-04 20:31:10 i see your dilemma jomat :/ 2014-10-04 20:32:50 jomat like barthalion wrote what about xmpp+otr 2014-10-04 20:33:45 but that's not really video eh? 2014-10-04 20:34:29 crow: We already use otr with irc and xmpp, but thats text only.. 2014-10-04 20:35:43 xmpp can audio and wiedeo (just not encrypted with otr) 2014-10-04 20:36:02 i hate to say it, i don't voice and video even much less are incredibly difficult to encrypt. the realtime thing introduces a lot of attack surface. 2014-10-04 20:41:11 crow: jitsi can use xmpp and otr, do you know other implementations, perhaps not in java? :-D 2014-10-04 20:42:36 not realy :/ 2014-10-04 20:42:40 HlArvEe5nyd5: i assume that could be one reason i can't find anything usable for it :-) 2014-10-04 20:45:52 i'm even looking into doing traffic analysis of cctv cameras over encrypted wifi - i guess you are aware of this research of mine... 2014-10-04 20:46:40 i guess the best you can do is run a vpn and connect the endpoints directly. 2014-10-04 20:46:54 everything else is as bad as this. 2014-10-04 20:48:30 i also played with this thought. I'd also get rid of all this nat pain- 2014-10-04 20:50:55 and then palava.tv inside this vpn. yo dawg %) 2014-10-04 21:03:29 webrtc, frmblgrmbl. but i shut up enough scepticism for today. 2014-10-04 21:05:24 :-D 2014-10-04 21:05:47 Fortunately it's tomorrow in 3 hours :-) 2014-10-04 21:22:39 palava says it uses encrpytion... but i didn't find any hint yet what they use... 2014-10-04 21:30:37 does it matter? it's in a browser. 2014-10-04 21:32:08 true that 2014-10-04 21:40:24 jomat: how safe is webrtc? 2014-10-04 21:40:44 I've seen the development branch of firefox has it for video telephony 2014-10-04 21:41:53 safety is no requriement for webrtc afaik 2014-10-04 21:42:11 also since it runs in the browser the question is kinda moot. :) 2014-10-04 21:51:45 browsers are not anymore for browsing html, but for selling ads and more services while externalising the development costs of multiplatform apps onto the users. webrtc is exactly such a technology that serves the interests silicon valley but not the browser users. 2014-10-05 11:27:22 k0r10n: there is no reason to explicitly add anything to $depends if a binary or library links to it 2014-10-05 11:31:13 k0r10n: things like depends="qt" isn't something you should be doing 2014-10-05 11:31:35 it's not Arch, abuild finds sodeps :) 2014-10-05 11:39:01 barthalion: ok, ty 2014-10-05 11:39:28 barthalion: for which aport? 2014-10-05 12:19:15 probably each you posted on mailing list 2014-10-05 12:19:36 I believe openexr links to ilmbase 2014-10-05 12:19:41 so it'll be detected 2014-10-05 12:19:55 ligcc dependency of ilbase will be detected as well 2014-10-05 12:20:29 I have no reason to believe that abuild would miss that grantlee links to qt 2014-10-05 12:20:48 and there is no need to add qt to depends of qimageblitz 2014-10-05 12:21:19 the rule of thumb is to explicitly define dependencies only if something doesn't link directly 2014-10-05 12:21:34 or isn't a binary at all, see python or bash scripts 2014-10-05 12:22:09 if your package links to n libraries, but it also provides some bash or ruby (whatever) script, you need to add it to deps 2014-10-05 12:22:27 k0r10n: ↑ 2014-10-05 12:26:44 barthalion: thank you for explain :) 2014-10-05 14:19:00 you're welcome 2014-10-05 20:46:36 n8@all 2014-10-06 04:52:38 moin 2014-10-06 04:55:12 evening 2014-10-06 07:19:23 newest kernel 3.14.19-r3 crashes XEN at boot while loading Dom0 => http://pastebin.com/AQcCwjv5 --- Is there any fix around / already reported anyone this problem? 2014-10-06 09:04:22 xen_roger: looks like there have been some issues with xen recently 2014-10-06 09:04:27 https://lkml.org/lkml/2014/8/29/317 2014-10-06 09:05:21 https://forums.grsecurity.net/viewtopic.php?f=3&t=4053 2014-10-06 11:11:11 xen_roger: are you able to test the vanilla kernel? 2014-10-06 15:09:27 ncopa - using vanilla kernel seems to work ... sorry for late response. had to build iso first. so I think the changes in grsec are the reason. you could just boot a alpine-xen.iso build with 3.14.19-r3 to see the problem. Hardware / VM doesn´t matter. 2014-10-06 15:20:35 ncopa: so error is related to grsec :( 2014-10-06 15:42:31 ncopa: vanilla is also still linux-vanilla-3.14.17-r0 - linux-grsec-3.14.17-r0 also works fine! 2014-10-07 05:12:06 ncopa: sir, could I provide you with some help about fixing the kernel related problem with xen and linux-grsec > 3.14.17-r0? 2014-10-07 06:02:29 morning 2014-10-07 06:02:58 xen_roger: it was vanilla kernel-3.14.19 you tested right? 2014-10-07 06:03:30 could you please verify that CONFIG_RANDOMIZE_BASE was the same in grsec kernel and vanilla kernel 2014-10-07 06:10:57 ncopa: no, 3.14.17 is still in repo 2014-10-07 06:12:19 CONFIG_RANDOMIZE_BASE is not set in this versions, also find out that in the past there were problems with CONFIG_RANDOMIZE_BASE_MAX_OFFSET if it was not =0x20000000 on xen. but it is. 2014-10-07 06:14:09 this is in the changelog: 2014-10-07 06:14:10 commit b477ddaafca9bb828a9c90b7ca890ff4f73571c7 2014-10-07 06:14:11 Date: Sun Sep 28 19:26:12 2014 -0400 2014-10-07 06:14:11 Author: Brad Spengler 2014-10-07 06:14:11 Backport fix from https://lkml.org/lkml/2014/8/29/317 2014-10-07 06:14:11 Problem is a boot crash with Xen due to KASLR, also observed 2014-10-07 06:14:12 by one of our users: 2014-10-07 06:14:14 https://forums.grsecurity.net/viewtopic.php?f=3&t=4053 2014-10-07 06:14:16 http://grsecurity.net/changelog-stable2.txt 2014-10-07 06:14:31 i kinda thought that would solve the CONFIG_RANDOMIZE_BASE_MAX_OFFSET thingy 2014-10-07 06:14:50 there is a new kernel now too, 3.14.20 2014-10-07 06:15:51 kernelconfig.x86_64:CONFIG_RANDOMIZE_BASE_MAX_OFFSET=0x20000000 2014-10-07 06:16:20 thats what i thought we had it set to? 2014-10-07 06:16:23 was also in linux-grsec 3.14.19-r3 2014-10-07 06:16:43 yes, so we probably still have the problem in 3.14.20 2014-10-07 06:17:01 ok... engine started ;) I will take a look at it after some coffee. so plz stand by an I will check this out 2014-10-07 06:17:11 i will try reproduce it too 2014-10-07 06:17:38 if i understand you correctly, you didnt try 3.14.19 vanilla kernel? 2014-10-07 06:18:46 after 3.14.17, in 3.14.18 or so I enabled the CONFIG_RELOCATABLE and CONFIG_RANDOMIZE_BASE 2014-10-07 06:18:49 i think that might be related 2014-10-07 06:19:11 it also broke .x86 2014-10-07 06:19:16 32 bit 2014-10-07 06:20:04 no I did not. at time of testing vanilla was still 3.14.17 2014-10-07 06:20:13 ok 2014-10-07 06:20:43 so i'd like to test the vnailla 3.14.19 with CONFIG_RELOCATABLE=y and CONFIG_RANDOMIZE_BASE=y and CONFIG_RANDOMIZE_BASE_MAX_OFFSET=0x20000000 2014-10-07 06:21:38 linux-vanilla-3.14.17-r0 << still the only version online ... after apk update 2014-10-07 06:21:51 is the 3.14.20 in main = 2014-10-07 06:23:18 wait my repo is outdated ... on website I see 3.14.19 2014-10-07 06:23:41 yes, edge/main should have linux-vanilla-3.14.19 2014-10-07 06:24:37 I get this when I try boot xen in side qemu/kvm: https://dpaste.de/f91Z 2014-10-07 06:24:45 I suppose that is different thing 2014-10-07 06:25:47 try to assign more than one core you fix that 2014-10-07 06:25:55 and yes it is diffrent 2014-10-07 06:27:34 I am trying to reproduce int qemu 2014-10-07 06:27:56 would be much easier to fix it if I could do that 2014-10-07 06:28:37 if you build an fresh alpine-xen.iso you got the problem ... also in kvm 2014-10-07 06:28:43 i bet we can make this problem go away by disabling CONFIG_RANDOMIZE_BASE= 2014-10-07 06:29:14 xen_roger: i think i just did 2014-10-07 06:29:27 is tehre some boot option to not make it reboot instantly? 2014-10-07 06:30:15 no... just happend on hardware also like in kvm at myside 2014-10-07 06:32:11 qemu-system-x86_64 -smp 4 --enable-kvm -m 2048 -curses ... 2014-10-07 06:32:22 i think i was able to reproduce it with the above 2014-10-07 06:33:17 would be nice if we could have it not reboot instantly 2014-10-07 06:33:32 ok.... you could attach serial console to see the error 2014-10-07 06:34:02 http://pastebin.com/AQcCwjv5 << it looks like that 2014-10-07 06:34:06 to xen yper visor or the dom0 kernel? 2014-10-07 06:34:15 to xen 2014-10-07 06:34:52 dom0 kernel is crashing at the moment xen tries to load it 2014-10-07 06:34:59 ok 2014-10-07 06:35:23 3.14.20-grsec is built too now 2014-10-07 06:35:27 I think I ll continue with that 2014-10-07 06:35:37 i'll build 3.14.20 banilla too 2014-10-07 06:35:39 why I cant see the new kernel 2014-10-07 06:36:04 3.14.20 is not yet pushed 2014-10-07 06:36:18 also not seeing 3.14.19 2014-10-07 06:36:25 3.0/main? 2014-10-07 06:36:32 just 3.14.17 - yes sir 2014-10-07 06:36:35 i dont think i pushed it to v3.0 yet 2014-10-07 06:36:59 ok ... that yould be the reason ;) 2014-10-07 06:37:06 i'd like to try fix those things before pushing anything to v3.0 2014-10-07 06:37:30 ok, 3.14.20 w/o xen boots in qemu 2014-10-07 06:37:49 could you check it into edge / testing ?= 2014-10-07 06:38:04 yes 2014-10-07 06:38:07 well 2014-10-07 06:38:11 edge/main 2014-10-07 06:38:19 that I can do now 2014-10-07 06:38:58 took me a half day after update to get my systems running again 2014-10-07 06:39:32 xen_roger: i'm sorry about that. it should not have happened 2014-10-07 06:39:37 what happened was 2014-10-07 06:39:41 morning 2014-10-07 06:39:54 they changed the config dependency, so for efi64 they required reolcatable kernel 2014-10-07 06:40:06 since we wanted efi, we got reolcatable 2014-10-07 06:40:15 anyway ... so I could talk with you ... that´s also nice ;) 2014-10-07 06:40:30 that triggered the "do you want randomize_base?" which i enabled 2014-10-07 06:40:43 i think maybe I should try disable that as next step 2014-10-07 06:40:47 and ship 3.0.5 2014-10-07 06:41:06 i think we might be able to fix this today 2014-10-07 06:41:08 so how did you fix it at the moment ? 2014-10-07 06:41:16 i havent fixed it 2014-10-07 06:41:44 i'd like to try 3.14.20, both vanilla and grsec 2014-10-07 06:41:55 if it happens with vanilla too then its upstream bug 2014-10-07 06:42:16 if it does not happen with vanilla, then i need report it to pipacs, the PaX developer 2014-10-07 06:42:47 if you want I could try disable randomize_base 2014-10-07 06:42:50 problems with xen? 2014-10-07 06:43:26 ScrumpyJack: yes 2014-10-07 06:43:33 if that fixes it i could simply disable that for v3.0 branch and try fix it for v3.1, without panic... 2014-10-07 06:44:28 ScrumpyJack: yes, poblems with xen 2014-10-07 06:49:59 xen_roger: is this our issue? https://dpaste.de/BBkG 2014-10-07 06:50:55 yes sir - looks like my error 2014-10-07 06:51:00 perfect 2014-10-07 06:51:03 i can reproduce then 2014-10-07 06:51:10 in a dev friendly env 2014-10-07 06:51:35 testing vanilla kernel now w/o xen 2014-10-07 06:51:35 +1 for kvm :) 2014-10-07 06:52:03 ok, 3.14.20 vanilla works without xen (just double check so we dont get fooled) 2014-10-07 06:52:13 now I'll try boot the xen with vanilla kernel 2014-10-07 06:53:24 ACTION crosses his fingers 2014-10-07 07:04:02 how can I tell alpine-iso to use edge for alpine-xen build ? 2014-10-07 07:24:21 back, sorry, had visitors in my office 2014-10-07 07:24:43 wb 2014-10-07 07:24:54 xen_roger: it will use whatever you have in /etc/apk/repositories 2014-10-07 07:25:03 it might be you can overrid APK_OPTS or something 2014-10-07 07:25:14 yes, found it on the wiki ... 2014-10-07 07:25:29 i can also confirm that xen+vnailla 3.14.20 *works* 2014-10-07 07:25:38 just added my local builds to it ... thats what i missed 2014-10-07 07:25:51 so i will have to talk with pipacs 2014-10-07 07:26:12 this is most likely an issue with PaX patch 2014-10-07 07:28:20 so xen + linux-grsec 3.14.20 does not work but vanilla does?! 2014-10-07 07:32:58 ncopa: thanks so far, my build is still running ... will try both versions my self an report 2014-10-07 07:47:11 confirm: xen + linux-grsec 3.14.20 => crash | xen + linux-vanilla 3.14.19 => crash 2014-10-07 07:47:25 huh? 2014-10-07 07:47:35 xen + linux-vanilla crash? 2014-10-07 07:47:40 i didnt do that for me 2014-10-07 07:48:07 3.14.20 still compling ... 2014-10-07 07:48:13 atleast not 3.14.20 2014-10-07 07:49:11 btw: when I change the kernel_flavour in alpine-xen.conf.mk it does not set the set it correct in syslinux.conf 2014-10-07 07:49:20 or isolinux.conf 2014-10-07 07:50:31 http://pastebin.com/XgVMyKiS << default line always points to xen-grsec 2014-10-07 07:50:56 ah 2014-10-07 07:51:11 did I forget to change something? 2014-10-07 07:51:11 KERNEL_FLAVOR := vanilla 2014-10-07 07:51:11 KERNEL_FLAVOR_DEFAULT := vanilla 2014-10-07 07:51:19 does that change anything? 2014-10-07 07:52:32 ... testing 2014-10-07 07:53:35 um 2014-10-07 07:53:47 i upgraded linux-vanilla to 3.14.20 2014-10-07 07:54:37 u r great! :) ty 2014-10-07 07:55:04 description is wron "Linux kernel with grsecurity" 2014-10-07 07:55:07 sure.. but now its cumbersome for me to test 3.14.19 vanilla... :) 2014-10-07 07:55:45 not working ... already did it but keep going 2014-10-07 07:57:10 ncopa: the description for the linux-vanilla package is wrong. should not be "Linux kernel with grsecurity" 2014-10-07 07:57:50 doh 2014-10-07 07:58:13 and yes ... KERNEL_FLAVOR_DEFAULT := vanilla solves the other little prob on my side 2014-10-07 08:03:02 and yes, linux-vanilla 3.14.20 works! puhhh... hope grsec version get fixed soon also. 2014-10-07 08:03:15 ok good! 2014-10-07 08:03:18 very well 2014-10-07 08:03:40 then we are 2 who came to the same conclusion 2014-10-07 08:03:50 yes sir 2014-10-07 08:03:59 mission complete - so far 2014-10-07 08:04:04 i havent verified that linux-vanilla 3.14.19 actually fails, but i believe it does 2014-10-07 08:04:11 i am waiting for pipacs to show up 2014-10-07 08:04:16 i have the info he asked for 2014-10-07 08:04:41 i think there is hope for fixing it relatively quick 2014-10-07 08:05:30 3.14.19 gives some more output and does not reboot but also bad type after mapping kernel into physical mem 2014-10-07 08:05:49 would be nice ... 2014-10-07 08:06:26 ncopa: could I pm you ? 2014-10-07 08:06:34 sure 2014-10-07 08:07:32 i'll go look for coffe 2014-10-07 08:07:43 ok 2014-10-07 08:11:17 xen_roger: you don't happen to know how to properly get Alpine running on Linode do you? 2014-10-07 08:11:57 no sry... never used 2014-10-07 08:12:30 m'kay no problem :) 2014-10-07 08:12:37 any recommended VPS hosting? 2014-10-07 08:12:41 but! ther is a guid replacing existing linux with alpine ... maybe not the best way but may work 2014-10-07 08:13:53 also, no sry... 2014-10-07 08:15:40 m'kay 2014-10-07 08:15:42 will try out 2014-10-07 08:16:00 systmkor1: http://wiki.alpinelinux.org/wiki/Replacing_non-Alpine_Linux_with_Alpine_remotely << if you not already found 2014-10-07 08:18:30 systmkor1: i have seen some alpine linux linode doc 2014-10-07 08:18:35 i dont remember where though 2014-10-07 08:18:51 ncopa: i found this http://uggedal.com/journal/alpine-linux-on-linode/ 2014-10-07 08:19:03 but it doesn't give much in the way of explaining the intermediate steps 2014-10-07 08:20:02 systmkor1: sounds easy ... did you try it? 2014-10-07 08:20:35 in the middle of but not sure were you go from grub command line to running a full shell 2014-10-07 08:22:49 as I see the script should bring up running alpine 2014-10-07 08:23:29 when I tried it just brings you to a grub command shell 2014-10-07 08:23:33 and that is it :\ 2014-10-07 08:23:40 will see if i goofed. 2014-10-07 08:24:40 does this shell brings all the stuff you need ? curl chroot etc. ? 2014-10-07 08:26:37 ah I think I found where I messed up 2014-10-07 08:26:38 one sec 2014-10-07 08:26:58 hope you got it ... 2014-10-07 08:27:30 oh also you guys know of anyone working on a solid docker file build script 2014-10-07 08:27:37 and defualt configurations 2014-10-07 08:27:39 ? 2014-10-07 08:28:02 yup cool booted into shell 2014-10-07 08:28:22 missed to go to rescue ? what was the problem ? 2014-10-07 08:28:54 systmkor1: i think uggedal was playing with docker 2014-10-07 08:28:57 what do you mean with docker build script ? 2014-10-07 08:29:29 well basically a well documented dockerfile 2014-10-07 08:29:36 and post docker image build scripts 2014-10-07 08:29:44 that get it up and running accordingly 2014-10-07 08:29:49 https://github.com/uggedal/docker-alpine 2014-10-07 08:29:54 wooooo 2014-10-07 08:29:57 thanks 2014-10-07 08:30:17 never used it 2014-10-07 08:30:35 ok 2014-10-07 08:30:39 good to know ncopa. would be nice to build alpine stuff ... 2014-10-07 08:30:57 i have got vmlinux, with xen problem 2014-10-07 08:31:59 ncopa: say what? 2014-10-07 08:37:19 http://dev.alpinelinux.org/~ncopa/xen/boot/ 2014-10-07 08:37:21 its for pipacs 2014-10-07 08:37:25 he asked for vmlinux 2014-10-07 08:37:55 i find it interesting that 3.14.19 vanilla does not work but 3.14.20 works 2014-10-07 08:38:49 yes... I am really interrested in the solution of this 2014-10-07 08:39:20 https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/?id=0c8fed7d9e198347cd9b5ece3dd0010dc507957d 2014-10-07 08:39:27 i think that was backported to grsecurity 3.14.19 2014-10-07 08:39:51 well I gotta crash and get some sleep 2014-10-07 08:39:56 thanks ttyl 2014-10-07 08:40:03 good night systmkor1 2014-10-07 08:40:06 ciao 2014-10-07 08:40:35 this looks like relevant: https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/?id=7fad1f4c34b4321a23946178415b1705f67af460 2014-10-07 08:46:08 sounds so... 2014-10-07 08:58:45 ok, I need help from pipacs with this 2014-10-07 10:51:39 morning all. anyone with any experience of osmocom on alpine? 2014-10-07 11:03:36 Guest3557: first time i hear about it 2014-10-07 11:05:41 ah, it's a GSM baseband software implementation 2014-10-07 11:06:26 what's the best way of installing texinfo4 for alpine - can I grab a tarball from gnu's ftp server, or should I recompile for alpine 2014-10-07 11:09:09 i'd say git lcone aports tree and cp -r texinfo texinfo4 and change pkgver in the APKBUILD 2014-10-07 11:09:22 why does not texinfo 5 work? 2014-10-07 11:10:27 the instructions that I'm following to build an gnu-arm toolchain say that gcc 4.5.2 won't work with texinfo5 2014-10-07 11:10:48 but they also don't say that gcc v4.9 will work, which would solve this particular issue 2014-10-07 11:12:46 added to this, the ftp servers where the source is held are a bit flaky and sometimes give corrupted downloads... so... :D 2014-10-07 11:16:14 ncopa, if you don't mind - the software also requires some dependencies that I cna't see in the alpine package list. mpfr4, for example 2014-10-07 11:16:50 if the dependencies aren't available, is this a pointer that I will really struggle to build this on alpine? 2014-10-07 11:17:04 yeah 2014-10-07 11:17:11 gcc-4.5.2 2014-10-07 11:17:25 sounds like you will have to struggle regardles... 2014-10-07 11:17:56 woot :P 2014-10-07 11:18:51 xen_roger: i got pipacs attention now and he is looking at it 2014-10-07 11:18:59 i also tried disable CONFIG_RANDOMIZE_BASE 2014-10-07 11:19:02 didnt make any difference 2014-10-07 16:29:05 ncopa: just notice yet, that the vanilla kernel has diffrent value CONFIG_RANDOMIZE_BASE_MAX_OFFSET=0x40000000 2014-10-07 16:29:58 where linux-grsec uses CONFIG_RANDOMIZE_BASE_MAX_OFFSET=0x20000000 2014-10-07 23:47:57 clandmeter, around ? 2014-10-08 06:07:28 ncopa, i pushed an uniso fix to alpine-conf.git 2014-10-08 06:07:38 i also pushed updated apk-tools to edge 2014-10-08 06:07:54 will start working on new apk-tools features soon 2014-10-08 06:09:26 cool! thanks! 2014-10-08 06:09:50 not sure if they are good for 3.0-stable yet, but at least for next edge tarball it'd be nice 2014-10-08 06:09:58 i was thinking something about apk-tools yesterday evening 2014-10-08 06:09:58 or edge snapshot that is 2014-10-08 06:10:08 but i dont remember what i was thinking now... 2014-10-08 06:10:20 :) 2014-10-08 06:11:13 i'm going to look at our redmine tickets #2251 and #3027 next 2014-10-08 06:11:42 i also plan to embed libfetch, and drop forking/exec of wget 2014-10-08 06:12:00 i'm ok with that 2014-10-08 06:12:09 few optional additions are merging post/pre install scripts 2014-10-08 06:12:19 seems that most upgrade/install scripts are just symlinks to each other 2014-10-08 06:12:51 and adding user/group creation 2014-10-08 06:13:41 i will try get xen working for 3.14.20 kernel 2014-10-08 06:13:46 so i can make 3.0.5 2014-10-08 06:14:12 right 2014-10-08 06:14:33 i think i want the two tickets + libfetch done. and if there's time then the script merging 2014-10-08 06:14:43 i had this idea for fetchers 2014-10-08 06:14:44 user/group creation stuff we probably can postpone for later on 2014-10-08 06:14:57 like /lib/apk/plugins/http.so 2014-10-08 06:15:04 and /lib/apk/plugins/ftp.so 2014-10-08 06:15:05 etc 2014-10-08 06:15:08 mmm 2014-10-08 06:15:13 libfetch does not work with that 2014-10-08 06:15:28 but probably not worth the administration coding for plugin system 2014-10-08 06:15:34 yeah 2014-10-08 06:15:41 and static build does not work with that anyways 2014-10-08 06:16:13 iirc, the libfetch code for each url handler is small enough anyway that the plugin system would have more overhead than benefit 2014-10-08 08:18:49 how to open luks drive at boot-time on alpine? like crypttab on other distro? 2014-10-08 08:19:02 not 100% someone sent me patch to support it 2014-10-08 08:19:34 you mean for encrypted root? 2014-10-08 08:19:45 or crypted /var or /usr or /home 2014-10-08 08:20:18 for encrypted root i think they added some boot option 2014-10-08 08:20:23 check http://git.alpinelinux.org/cgit/mkinitfs/tree/initramfs-init.in 2014-10-08 08:20:23 booting from usb, using lbu and storing data on disk with luks 2014-10-08 08:20:33 hm ok 2014-10-08 08:20:47 so its not from initfs 2014-10-08 08:21:14 i am not sure we have anything specific for it 2014-10-08 08:21:22 i dont think we have 2014-10-08 08:21:52 we could check what gentoo does, but i suspect we will want to have add support for a crypttab 2014-10-08 08:22:10 on debian or something I just add it to fstab and crypttab and it get mounted on boot. 2014-10-08 08:22:35 yeah 2014-10-08 08:22:46 i dont think its too difficult to implement 2014-10-08 08:22:49 ok... would be nice to store sensible date crypted on disk while booting from live media 2014-10-08 08:22:56 yes agree 2014-10-08 08:25:37 xen_roger: it woudl be great if you could have a look at how to get that working nicely 2014-10-08 08:27:12 I will take a look at it... but first i need to fix my xen systems after testing this patch ... thank you btw. for fast helping with this 2014-10-08 08:28:14 sorry for breaking it in the first place :-( 2014-10-08 08:28:27 sh*t happens ;) 2014-10-08 08:29:11 unfortunally it does 2014-10-08 08:29:38 yeah! it boots up 2014-10-08 08:30:04 did you rebuild it yourself? 2014-10-08 08:30:13 yes sir 2014-10-08 08:30:15 i think i messed up the commit 2014-10-08 08:30:20 ok that explains :) 2014-10-08 08:30:31 i just pushed checksum fix 2014-10-08 08:30:39 too much happening at same time :-( 2014-10-08 08:31:03 np 2014-10-08 08:34:40 Hmm, where's the best place to file a bug for /usr/share/mkinitfs/initramfs-init? 2014-10-08 08:37:54 bug tracker or here is seems a goot point to start 2014-10-08 08:38:45 jomat, http://bugs.alpinelinux.org/projects/alpine/issues/new set category to 'Boot sequence' 2014-10-08 08:41:29 thx 2014-10-08 10:34:52 how can I commit stuff to git of alpine ? 2014-10-08 10:42:33 you can only commit in your local clone 2014-10-08 10:43:02 to be able to git push it to git.alpinelinux.org you need permission (me adding your ssh key to there) 2014-10-08 10:44:56 xen_roger: git add APKBUILD dmcrypt.initd dmcrypt.confd && git commit 2014-10-08 10:45:06 as commit message you use 2014-10-08 10:45:34 main/cryptsetup: add init script 2014-10-08 10:46:02 then to show patch here you can do: git format-patch -1 --stdout | sprunge 2014-10-08 10:46:32 or you can git send-email --to alpine-devel@lists.alpinelinux.org -1 2014-10-08 10:46:50 looking to send it to mailing list ... but unsure about commit comment 2014-10-08 10:47:13 first line in commit message is 'subject' 2014-10-08 10:47:19 then blank line as separator 2014-10-08 10:47:34 then comes the 'body' for the email 2014-10-08 10:48:05 git format-patch -1 --stdout 2014-10-08 10:48:10 done some mistake ... can I reset commit to add new subject ? 2014-10-08 10:48:10 will show you how it looks 2014-10-08 10:48:17 git commit --amend 2014-10-08 10:48:27 will let you rewrite commit message 2014-10-08 10:49:32 is that ok ? - [PATCH] main/cryptsetup add: support for dmcrypt at boot 2014-10-08 10:50:35 main/cryptsetup: add support for dmcrypt at boot 2014-10-08 10:50:43 I see 2014-10-08 10:50:46 is perfect 2014-10-08 10:51:23 if you had longer explanatino on why it was needed et you could have added that too 2014-10-08 10:51:26 but i dont think its needed 2014-10-08 10:51:38 you can have a look at how other commits are done: http://git.alpinelinux.org/cgit/aports/log/ 2014-10-08 10:51:53 how to configure git sendmail ? 2014-10-08 10:52:24 sry never used it 2014-10-08 10:53:14 i found this useful when setting up git send-email: https://coderwall.com/p/dp-gka 2014-10-08 10:54:21 http://wiki.alpinelinux.org/wiki/Creating_patches 2014-10-08 10:54:30 http://wiki.alpinelinux.org/wiki/Development_using_git 2014-10-08 10:56:54 all done but git-send-email --- still cry about 127.0.0.1 o.o 2014-10-08 10:57:38 git config --global sendemail.smtpserver smtp.gmail.com 2014-10-08 10:57:59 you need set which smtpserver to use in your ~/.gitconfig 2014-10-08 10:58:17 smtpserver = smtp.gmail.com 2014-10-08 10:59:06 that´s what I forgot ... thx 2014-10-08 11:16:07 ncopa: after mess around with fw + mailer; done! 2014-10-08 11:23:07 xen_roger: if i understand that right, you configure your stuff in /etc/conf.d/dmcrypt 2014-10-08 11:23:18 and does not support /etc/crypttab 2014-10-08 11:23:33 yes sir ... its ported from how gentoo does 2014-10-08 11:24:08 so I use it on my gentoo too 2014-10-08 11:24:08 are you sure that is what we want to do? 2014-10-08 11:24:12 ok 2014-10-08 11:24:28 do you have a doc for how to configure conf.d in gentoo? 2014-10-08 11:24:40 gentoo + openrc = alpine + openrc => so why not? 2014-10-08 11:26:34 like this ? 198.252.153.112 2014-10-08 11:26:36 sry 2014-10-08 11:26:50 this https://www.gentoo.org/doc/en/handbook/handbook-x86.xml?part=2&chap=4#doc_chap4 2014-10-08 11:28:07 thats about writing init.d scripts 2014-10-08 11:28:22 section 4.c. 2014-10-08 11:28:35 what exactly do you want to know ? 2014-10-08 11:29:03 i wasnted some examlple how to set up the conf.d/dmcrypt 2014-10-08 11:29:09 but that file has examples 2014-10-08 11:29:12 i can look at those 2014-10-08 11:29:26 yes file is full of use-cases 2014-10-08 11:29:51 missunderstanding ... ;) 2014-10-08 11:31:52 works fine on my side ... so would be nice if we could add it 2014-10-08 11:32:09 i just have the feeling that it is a bit hackish 2014-10-08 11:32:23 i dont doubt that it works or not 2014-10-08 11:32:32 i am in doubt if i want maintain it 2014-10-08 11:33:04 or if we should got for /etc/crypttab 2014-10-08 11:33:36 eg debian style instead of gentoo style 2014-10-08 11:35:47 did you look at how debian it does ? 2014-10-08 11:35:54 not really 2014-10-08 11:36:02 have you? 2014-10-08 11:36:05 looks like a lot more ... ;) 2014-10-08 11:36:51 in any case, i suppose i can add this init.d script 2014-10-08 11:38:49 just one of the scripts that get called http://paste.debian.net/125132/ 2014-10-08 11:39:29 ugh 2014-10-08 11:39:34 i see what you mean yes 2014-10-08 11:39:35 wich is called by http://paste.debian.net/125134/ 2014-10-08 11:39:38 :) 2014-10-08 11:40:14 i love debian also ... but I think the gentoo way here is much more easier 2014-10-08 11:40:30 the /etc/conf.d/dmcrypt instead of /etc/crypttab format? 2014-10-08 11:40:31 ok 2014-10-08 11:40:49 if you have done both and think gentoo way is better 2014-10-08 11:40:52 then lets go there 2014-10-08 11:42:04 i wil rename dmcrypt.rc to dmcrypt.initd so it gets consistent with the other scripts 2014-10-08 11:42:15 gentoo way are just 2 files ... and I use it on my gentoo this way with lvm on luks on raid5 2014-10-08 11:42:27 ok feel free 2014-10-08 11:44:10 the pkgrel needs increase too or builder will not rebuild it, and apk upgrade will not see that it needs update 2014-10-08 11:44:12 but i fix that too 2014-10-08 11:44:56 ok. my fail ... next time 2014-10-08 11:45:13 np, it was a pretty good first timer :) 2014-10-08 11:45:42 the libs part also is unchanged ... could be removed from patch 2014-10-08 11:53:43 xen_roger: thanks alot 2014-10-08 11:54:22 you are welcome ... 2014-10-08 11:54:24 ;) 2014-10-08 12:14:25 rnalrd: are you able to update linux-virt-grsec in 3.0-stable? 2014-10-08 12:14:31 or can i just cherry pick it? 2014-10-08 12:14:37 sure 2014-10-08 12:14:55 thanks 2014-10-08 12:16:52 re ceph, musl res_* are threadsafe and stateless 2014-10-08 12:17:07 which means we can probably just use res_query instead of res_nquery 2014-10-08 12:17:23 code will likely need to be ported 2014-10-08 12:20:19 rnalrd: i think that is the stopper for v3.0.5 atm 2014-10-08 12:20:39 k, i'll do right away 2014-10-08 12:20:48 super. thanks 2014-10-08 19:44:53 does anyone know the reason for: "WARNING: No provider for dependency 'uclibc-utils'" while building an iso with alpine-iso? 2014-10-08 19:45:52 im not sure 2014-10-08 19:45:57 but could be initramfs config? 2014-10-08 19:46:12 check in etc if you have new configs around 2014-10-08 19:47:53 sry, don´t understand. it´s clean fresh clones alpine-iso und just running e.g "make PROFILE=alpine-mini iso" and get the error above ... other images the same ... 2014-10-08 19:48:16 its only a WARNING and does not broke the image, but wonder why 2014-10-08 19:48:17 evening gentlemen 2014-10-08 19:49:18 xen_roger: is this for 2.7 or 3.0? 2014-10-08 19:49:27 3.0 2014-10-08 19:50:09 i didnt get that error, and i build one pretty recent 2014-10-08 19:50:20 i've not been following the chatter. I'm in need of some alpine-xen magic. Can I use 3.0.5 in production? 2014-10-08 19:50:46 losely speaking 2014-10-08 19:50:57 xen_roger: i did have errors previously, but i had some older configs around which needed updating. 2014-10-08 19:51:14 ScrumpyJack: after the last update -> 3.0.5 I think yes ... ;) 2014-10-08 19:51:34 are you sure you cloned master alpine-iso? 2014-10-08 19:51:50 clandmeter: yes sir, fresh clone 2014-10-08 19:52:40 xen_roger: I appreciate your speedy response, but the use of the verb "think" worries me :) 2014-10-08 19:52:58 what is the overall level of confidence in alpine-xen 3.0.5 ? 2014-10-08 19:53:35 ScrumpyJack: what means production? 2014-10-08 19:55:32 alpine-xen is my prefered dom0 setup. non-commercial production 2014-10-08 19:55:40 xen_roger: what does: find /etc -iname *apk-new 2014-10-08 19:55:58 ScrumpyJack: it is also on my side 2014-10-08 19:56:34 and yes you can go with 3.0.5 because, prevously bugs in dom0 got fixed! now 2014-10-08 19:57:17 xen_roger: ooo, cool. what do you use for manage disks and what filesystem do you use on the disks? 2014-10-08 19:58:36 clandmeter: just a few not important ones like, hosts, shadow, keymaps group fstab passwd and hostname ... do you mean one of this is the reason? 2014-10-08 20:00:29 ScrumpyJack: I am using ext4 on dom0 and inside domu ... currently after my patch was acepted I migrate my domu into a luks container an thinking about switching from raw - full-provision images to lvm 2014-10-08 20:02:11 xen_roger: yes, I already use LVM + ext4. I want to use btrfs now. 2014-10-08 20:02:28 btrfs because of? 2014-10-08 20:02:43 xen_roger: no, only ones from /etc/mkinitfs/features.d 2014-10-08 20:03:18 xen_roger: clones and snapshots with LXC 2014-10-08 20:03:21 ACTION want to move away from btrfs 2014-10-08 20:03:27 clandmeter: maybe I take a look at it another time... 2014-10-08 20:04:00 ScrumpyJack: isn´t it also possible with lvm? 2014-10-08 20:04:33 clandmeter: interesting. can you elaborate? I've not heard many opinions 2014-10-08 20:04:48 ScrumpyJack: planning to migrate to zfs 2014-10-08 20:04:56 since its working now on alpine. 2014-10-08 20:05:27 ah zfs, yes, that is the holy grail, but i don't have that many disks 2014-10-08 20:05:36 im missing swap support on btrfs 2014-10-08 20:05:59 and zfs just seems more logical then any other fs. 2014-10-08 20:06:21 xen_roger: yes, you can do it with LVM for xen, but btrfs works well with LXC, or so I have read 2014-10-08 20:07:05 on my lxc i have the feeling btrfs is slower. 2014-10-08 20:07:19 but didnt really do any tests (just feeling) 2014-10-08 20:07:31 is zfs in alpine-xen? how has it been implemented? 2014-10-08 20:07:58 dunno, i dont use xen 2014-10-08 20:08:13 is zfs in alpine ready to use? 2014-10-08 20:08:20 yes 2014-10-08 20:08:35 i have a test setup 2014-10-08 20:08:41 although not heavily tested 2014-10-08 20:08:45 ScrumpyJack: are you using live-disk ? 2014-10-08 20:09:12 ive build alpine-zfs iso, so i can run from ram with zfs. 2014-10-08 20:09:27 thats why I ask 2014-10-08 20:09:56 clandmeter: cool. although that can't be distributed :( 2014-10-08 20:09:56 any doc about alpine zfs ? 2014-10-08 20:10:04 its nothing special, just include zfs and spl in initramfs 2014-10-08 20:10:29 xen_roger: live-disk? 2014-10-08 20:10:57 err not initramfs, modloop 2014-10-08 20:11:02 xen_roger: MODLOOP_EXTRA := zfs-grsec spl-grsec 2014-10-08 20:11:13 add that to your alpine-iso conf 2014-10-08 20:11:14 live-disk means running from ram ... disk/usb/sd what ever 2014-10-08 20:13:30 --> The only major drawback is that native ZFS encryption is not available ... :( 2014-10-08 20:14:40 on my dev box, i load it from a usb stick and the xl xen guests are on the disks. It works really well. I back up the apkvol regularly 2014-10-08 20:16:28 using this way in prouction ... currently I am going to encrypt the disks domus are on 2014-10-08 20:16:46 xen_roger: can i ask why? 2014-10-08 20:17:26 to protect sensible data inside domu 2014-10-08 20:17:48 why else ? 2014-10-08 20:18:58 sorry, i didn't mean to be awkward 2014-10-08 20:19:39 no no ... 2014-10-08 20:20:20 the other way would be crypto-layer inside domu wich does not benefit from hardware crypto support of cpu I think 2014-10-08 20:20:21 is this to help protect data if your disks are removed? 2014-10-08 20:22:04 it protects agains raids ;) disk rma or every other offline access to the data without permit 2014-10-08 20:23:02 yeah. raids are a problem. what with all the vikings around :) :) 2014-10-08 20:23:23 amyway, thanks for the info on alpine-xen. I'm looking forward to using it 2014-10-08 20:23:33 so, yes - pull out some disk and data on it is useless even not able to know about wich data was on it 2014-10-08 20:26:02 ScrumpyJack: could you give me an example of lvm disk line in you config? 2014-10-08 20:30:37 just beeing confused with the syntax ... but I think I got it 2014-10-08 20:31:45 arg, i can send it in the morning. the dev host is at work :( 2014-10-08 20:34:49 I updated the alpine wiki virtualisation sections, but i think without LVM. I should add that. I've learnt a lot with alpine about xen and lxc since then 2014-10-08 20:36:55 http://wiki.alpinelinux.org/wiki/Tutorials_and_Howtos#Virtualization 2014-10-09 06:46:03 morning 2014-10-09 08:47:57 i can't seem to pull the latest git repo. This is a problem my end, but im not sure how to fix it. 2014-10-09 08:48:16 error message? 2014-10-09 08:50:17 http://sprunge.us/FUAe 2014-10-09 08:51:01 how can I tell what is unmerged? 2014-10-09 08:51:30 git status 2014-10-09 08:51:54 yay! 2014-10-09 08:52:11 but the list is the long list in the sprunge 2014-10-09 08:52:43 git status | sprunge 2014-10-09 08:53:21 git status showed three files in red that I added with git add 2014-10-09 08:53:35 ok 2014-10-09 08:53:44 you should probably resolve the conflicts first 2014-10-09 08:53:56 check if there are any >>>>>>> in the rile 2014-10-09 08:53:57 file* 2014-10-09 08:54:00 I have added them, but of course I can't commit them 2014-10-09 08:56:35 how about git reset --hard HEAD ? 2014-10-09 08:56:47 you'll loose changes 2014-10-09 08:56:53 lose* 2014-10-09 08:57:15 I don't think there is anything I want to keep, im sure you committed everything to testing a while ago. I've not made changes since then 2014-10-09 08:57:17 so if you have done something different than from git.a.o you'll lose your changes 2014-10-09 08:57:41 i think you get this problme becuase i changed your commits 2014-10-09 08:57:51 so you can probably just do git reset --hard HEAD 2014-10-09 08:57:57 ah ok cool. I'll do that 2014-10-09 08:58:00 you might need to do it twice 2014-10-09 08:58:56 I'm up to date now. \o/ thanks for your help 2014-10-09 11:34:17 kaniini: do you know if the websockets patch set is upstreamed in xen-4.4? 2014-10-09 11:34:23 qemu-xen-websocket.patch 2014-10-09 11:34:23 qemu-xen-tls-websockets.patch 2014-10-09 11:34:23 qemu-coroutine-gthread.patch 2014-10-09 11:34:23 qemu-xen-websocket-plain-hack.patch 2014-10-09 11:34:23 qemu-xen-vnc-robustness.patch 2014-10-09 11:42:48 kaniini: why do we want gthread for coroutines? should we do that for the 'qemu' app too? 2014-10-09 11:53:56 it's not 2014-10-09 11:54:08 and, the coroutines support was broken 2014-10-09 11:58:44 websockets appears to be paritially upstreamed 2014-10-09 11:59:30 of the above patches i keep qemu-coroutine-gthread.patch and remove the rest 2014-10-09 11:59:39 ok? 2014-10-09 11:59:51 it is for xen-4.4 in edge 2014-10-09 12:02:00 i think he mentioned to abandoned xen for vmware 2014-10-09 12:02:50 iirc he said his new employee, which does not need to mean that himself still use xen 2014-10-09 12:02:58 or want continue maintain xen 2014-10-09 12:04:21 its long time ago, but it sounded he was not interested in xen much. thats why he didnt fix hvmloader. 2014-10-09 12:05:02 but anyway, he's probably the best source to provide you that information. 2014-10-09 13:05:36 royger: i'm looking at splitting out the python bits from xen and found those python scripts: 2014-10-09 13:05:40 http://sprunge.us/LDSQ 2014-10-09 13:05:56 are there some logic way to put thsoe in subpackages? 2014-10-09 13:06:15 or are xen completely useless without thsoe tools? 2014-10-09 13:27:50 ncopa: 2014-10-09 13:28:08 ncopa: the only one that seems relevant is pygrub 2014-10-09 13:28:22 the rest are debugging tools or outdated utilities AFAICT 2014-10-09 13:29:24 if you want to strip them all, Xen would still be functional, but in order to boot PV DomU you would have to use pvgrub or have the kernel in the Dom0 and use the "kernel" and "initramfs" options in the config file 2014-10-09 13:30:29 ncopa: BTW, I've asked about rewritting pygrub in C, but the response was that pygrub is a security critial component, so the new C code should be audited very deeply, and there's no guarantee it would be added 2014-10-09 13:31:05 yes but with pvgrub, tehre is no need for pygrub right? 2014-10-09 13:33:04 IMHO, I think so, but I'm not sure pvgrub can boot everything pygrub supports 2014-10-09 13:34:07 I would pack all this stuff into something like xen-python-tools and let each user decide. In the long run is better to prevent users from using pygrub, pvgrub is much more secure (because it runs on guests context) 2014-10-09 13:34:53 i have pvgrub booting lots of stuff 2014-10-09 13:35:13 i find it much better than pygrub 2014-10-09 13:35:55 lemme check 2014-10-09 13:36:59 ScrumpyJack, ncopa: we could also provide pvgrub2, which should probably be better: http://lists.gnu.org/archive/html/grub-devel/2013-11/msg00085.html 2014-10-09 13:37:39 the thing with pvgrub is that you need to know the guest bitness (i386 or amd64) before botting it 2014-10-09 13:38:26 ok 2014-10-09 13:38:27 my success with pvgrub is sles11, opensuse debian and alpine 2014-10-09 13:38:49 i think what i'll do is, split out py-xen 2014-10-09 13:39:02 with the python libs and all the python based tools 2014-10-09 13:39:12 including py-xen 2014-10-09 13:39:22 i started with pygrub, but I hate python errors+debugging when things don't work 2014-10-09 13:39:33 so if you need pygrub you'll have to manually add py-xen 2014-10-09 13:39:47 this makes me very happy 2014-10-09 13:39:50 ScrumpyJack: I think pygrub is also capable of booting Solaris and NetBSD guests, but I have never tried it 2014-10-09 13:40:09 i suppose you can load netbsd and solaris with hvmloader 2014-10-09 13:40:18 ncopa: not if they are pv guests 2014-10-09 13:40:36 Solaris had PV support, NetBSD has PV support 2014-10-09 13:40:42 ok 2014-10-09 13:40:51 if i can't use pvgrub, i'd rather pass the the kernel and ramdisk files directly to xen 2014-10-09 13:41:20 that's how much i dislike pygrub :) 2014-10-09 13:41:56 isn't xenproject giving up support for all the python stuff anyway? 2014-10-09 13:42:05 ScrumpyJack: that's what I usually do, but this is not posible if you are a service provider and what users to be able to upgrade their kernels 2014-10-09 13:42:39 ScrumpyJack: only xend(xm) so far, but there are not many other python things 2014-10-09 13:44:27 ah yes, xend. that's dead now right? nothing wrong with xl :) 2014-10-09 13:46:13 and agreed about the kernel upgrade thing, although you may be able to provide each user with a filesystem to place a new kernel outside their guest that is loaded by xen. A bit of a faf admittedy. 2014-10-09 13:47:00 anyway, sorry, that's off topic :) 2014-10-09 13:49:35 ScrumpyJack: yes, xend has already been removed from current, next Xen version (4.5) will ship without it 2014-10-09 13:50:02 i will not build xend with xen-4.4 2014-10-09 13:50:25 better, xend has not been maintained for a long time 2014-10-09 13:50:58 it was just a big piece of crap IMO 2014-10-09 13:54:06 ok, after that py-xen subpackage, i think only thing left is that perl dep 2014-10-09 13:54:14 oh there are a bunch of bash scripts too 2014-10-09 13:54:23 would be nice port those to posix shell 2014-10-09 13:58:21 ncopa: I've tried to get some GSOC or OPW students to rework hotplug scripts, but so far no luck (I don't blame them, the project is not interesting at all) 2014-10-09 14:00:03 i want to use alpine-xen on a server (this will be my first time, all other alpine-xen's are on dev boxes) I think i'll wait for the python-free build. would you agree? 2014-10-09 14:00:20 *shrug* 2014-10-09 14:00:39 i just set up an alpine-xen-3.0.5 box 2014-10-09 14:00:44 pulls in python 2014-10-09 14:00:48 server = production (and I'm really excited to use alpine-xen) :) 2014-10-09 14:01:00 but the only damage that does is eating ram 2014-10-09 14:01:06 wasting ram 2014-10-09 14:34:04 and python is 36Megs! yuk! 2014-10-09 14:45:50 i think i managed to get rid pf perl too 2014-10-09 14:45:54 of* 2014-10-09 14:46:16 this is the hunk that uses perl: https://gist.github.com/ncopa/ab91578e59be530d9579 2014-10-09 14:46:35 the comment indicates that something is wrong with the locking strategy 2014-10-09 14:48:17 if they are not guaranteed that the stat(2) will lose the race 2014-10-09 14:48:25 and then they say 'perl is very fast' 2014-10-09 14:48:34 http://wiki.xen.org/wiki/PvGrub#Build 2014-10-09 14:48:45 what guarantee they have that forking perl wont lose the race? 2014-10-09 14:48:55 pvgrub seems to depend on python-dev in that link 2014-10-09 14:49:31 thats just a build time dep i suppose 2014-10-09 14:51:09 i hope so 2014-10-09 14:55:46 its a race between with-lock-ex and flock? 2014-10-09 14:55:55 dunno 2014-10-09 14:56:06 but the comment makes it sound that something is wrong 2014-10-09 14:56:52 pvgrub is just a kernel, it doesn't depend on anything 2014-10-09 14:57:00 *run-time wise 2014-10-09 14:58:22 royger: i think this C code should replace the perl snippet: https://gist.github.com/ncopa/bf50891994a16f3043ac 2014-10-09 14:58:39 we could probably install it in the LIBEXEC dir 2014-10-09 14:59:06 it should do the exact same thing as the perl snippet 2014-10-09 15:00:08 i wonder if they could simplify it a bit by using mkdir/rmdir tricks 2014-10-09 15:00:35 mm, this is not something I'm familiar with, could you send it upstream? 2014-10-09 15:00:48 i suppose i should 2014-10-09 15:00:54 if you hurry there's a change it will get into Xen 4.5 2014-10-09 15:01:16 i'd probably need to clean it up and integrate it with the build system 2014-10-09 15:01:28 i suppose i could post it as an RFC 2014-10-09 15:10:38 how much resource does alpine-xen use ? have small pc pc-engine apu4c wonder if i could use it for that 2014-10-09 15:10:51 esxi takes almost 50% of everything.. 2014-10-09 15:11:43 does this suggest that perl doesn't look at /proc? 2014-10-09 15:12:43 correct 2014-10-09 15:14:10 it just runs fstat on the given fd and stat on the given filename 2014-10-09 15:14:14 and compare the inode 2014-10-09 15:14:20 just like my C app does 2014-10-09 15:14:55 crow you could try :) 2014-10-09 15:15:19 ScrumpyJack well its "production" server, so if i try i need to get it working very soon.. 2014-10-09 15:15:38 ScrumpyJack or there is alpine-xen for run-from-ram to? 2014-10-09 15:18:01 yes. that's how a few of us do it here. alpine-xen on a USB stick that boots into ROM 2014-10-09 15:21:54 where do you store vms? 2014-10-09 15:22:08 on disks :) 2014-10-09 15:22:50 ScrumpyJack :) 2014-10-09 15:23:23 the disks are managed by LVM 2014-10-09 15:23:23 i mean on same USB disk , also where the lbu store the stuff, or some other media 2014-10-09 15:24:28 the apkvol is on the USB, and a backup is kept offsite over ssh 2014-10-09 15:25:15 like this ... http://wiki.alpinelinux.org/wiki/Alpine_local_backup#Use_SSH_to_create_an_apkovl_on_a_different_host 2014-10-09 15:27:52 ScrumpyJack this APU4C have an m-SATA SSD, maybe with some normal price m-sata ssd i could make it datastore 2014-10-09 15:28:22 ScrumpyJack i saw you wrote that you are preparing your production server for that, i hope you will write some Note and put them on wiki :D 2014-10-09 15:30:04 yes, i should do that, but i'm going to wait until ncopa has worked his magic and removed the python and perl dependancies from alpine-xen 3.0.6 :P 2014-10-09 15:31:41 ok 2014-10-09 15:34:12 crow: how long have you got? 2014-10-09 19:13:01 hey guys 2014-10-09 19:17:36 i'm a newbie and i want to join the dev 2014-10-09 20:00:06 ScrumpyJack i dont have m-sata ssd, but i could plug in 2014-10-09 20:11:27 sporgj: patches welcome 2014-10-09 21:48:12 barthalion: I don't even know where to start 2014-10-09 22:00:42 sporgj: check our wiki wiki.alpinelinux.org 2014-10-09 22:00:59 sporgj: download the latest edge iso and isntall it. 2014-10-09 22:01:06 sporgj: then install alpine-sdk 2014-10-09 22:01:20 sporgj: and clone our aports from git.a.o 2014-10-09 22:02:50 clandmeter: is it ok to work from a VM ? 2014-10-09 22:12:10 sporgj: sure 2014-10-09 22:12:32 i have my dev on lxc 2014-10-09 22:14:10 clandmeter: had never heard of that 2014-10-09 22:14:20 clandmeter: looks interesting 2014-10-09 22:14:37 it is :) 2014-10-09 22:15:00 any recent kernel should support it. 2014-10-09 22:15:11 its like a chroot on steroids 2014-10-09 22:17:09 i would give it a try 2014-10-09 22:17:20 keep in mind, I just got into kernel stuff this summer 2014-10-09 22:17:36 I've become obsessed :) 2014-10-09 22:20:59 ncopa: ive added opus support to ffmpeg (as you see) so i put opus directly into main. 2014-10-09 22:21:51 barthalion: any reason we are holding back on ffmpeg 2.4? i guess abi changes? 2014-10-10 01:15:58 clandmeter, there is /testing/opus 2014-10-10 04:11:44 clandmeter: API hasn't changed that much 2014-10-10 04:11:54 almost everything should be building 2014-10-10 04:12:12 I just didn't have motivation to take care of it here 2014-10-10 04:39:06 ah, yes, vlc has issues 2014-10-10 04:39:14 so we would need to switch to git 2014-10-10 04:39:35 because hacking autotools isn't enough and rmvb support is broken 2014-10-10 06:20:52 how can I see wich files are inside a *.apk package ? 2014-10-10 06:22:33 ok, got it ... by using "apk -L " 2014-10-10 07:45:34 morning 2014-10-10 07:50:38 mornin 2014-10-10 07:59:49 <[xming]> moin 2014-10-10 08:17:16 ncopa: could you have a look at http://bugs.alpinelinux.org/issues/3275 ? 2014-10-10 08:54:00 ScrumpyJack: whitespace damage, but i can fix it 2014-10-10 08:54:02 sed -E -i -e 's/\s+$//' APKBUILD 2014-10-10 08:54:22 whitespace before EOL 2014-10-10 09:09:13 ScrumpyJack: i think we want fltk built with system libs for jpeg and png 2014-10-10 09:09:18 and i think we want enable shared libs 2014-10-10 09:09:31 we probably also want xft and xext support? 2014-10-10 09:15:21 ok, I'll add that 2014-10-10 09:16:12 i'm on it 2014-10-10 09:16:17 do we want gl support? 2014-10-10 09:21:26 ncopa, mosquitto v1.3.5 is out, but may be wait for 1.4 2014-10-10 09:21:33 http://jpmens.net/2014/07/03/the-mosquitto-mqtt-broker-gets-websockets-support/ 2014-10-10 09:21:40 ! 2014-10-10 09:21:41 http://goochgooch.wordpress.com/2014/08/01/building-mosquitto-1-4/ 2014-10-10 09:21:42 \o/ 2014-10-10 09:21:47 :-)) 2014-10-10 09:57:30 ncopa thanks 2014-10-10 10:17:24 <[xming]> !help 2014-10-10 10:18:01 <[xming]> no bot responding :( 2014-10-10 10:21:49 we have humans to help. 2014-10-10 10:35:51 mmmm humans 2014-10-10 10:37:20 ACTION is on a boring conf call, it's been two hours, and feels like a zombie 2014-10-10 10:52:29 the phone is frying my brain 2014-10-10 11:07:03 [xming]: wanna help set up/maintain a proper bot? 2014-10-10 11:17:20 <[xming]> ncopa: yeah 2014-10-10 11:18:05 <[xming]> ncopa: probably a good time for me to write one 2014-10-10 11:24:13 <[xming]> ncopa: wht do you have in mind? 2014-10-10 11:32:28 i had in mind: "it would be nice if someone else could do it..." :) 2014-10-10 11:32:54 you seemed to have some expectation with !help 2014-10-10 11:39:33 <[xming]> maybe some lua bot, I have manged some supybots and others in the passed, but there is always something in those bots which I didn't like always wanted "my own bot" :p 2014-10-10 11:41:04 i like lua 2014-10-10 11:41:05 i 2014-10-10 11:41:13 i'd like be able to add some hooks too 2014-10-10 11:41:34 or have the bot listen to mqtt messages or so 2014-10-10 11:41:42 (for git commits etc) 2014-10-10 11:49:41 <[xming]> some bots have built-in http server, to which you can use post commit script to POST 2014-10-10 11:50:27 <[xming]> I haven't managed bots for many years, trying to see what/if new/better ones are out there 2014-10-10 11:54:51 ScrumpyJack: btw, re #3275 2014-10-10 11:55:14 what you did there is exactly the kind of help i want more of 2014-10-10 11:55:20 big thanks 2014-10-10 11:59:14 <[xming]> IIRC there are bots with redmine plugins 2014-10-10 11:59:31 <[xming]> so you say #3275 and it will print the bug's title 2014-10-10 11:59:52 <[xming]> I need to find one that suits this project and with miminal effort 2014-10-10 12:00:07 <[xming]> haven't found any lua bot that is advanced enough 2014-10-10 12:10:39 lua tends to be: do it yourself 2014-10-10 12:26:38 [xming]: might be we coudl use https://github.com/JakobOvrum/LuaIRC 2014-10-10 12:31:07 looks good 2014-10-10 12:31:59 <[xming]> I don't see a way to push info to it, so it will have to pull the git for commits 2014-10-10 12:33:22 push info? 2014-10-10 12:33:40 oh 2014-10-10 12:33:43 hm 2014-10-10 12:34:05 <[xming]> your git post commit hook 2014-10-10 12:34:11 yeah yea 2014-10-10 12:34:26 i suppose you could implement a unix socket listener... 2014-10-10 12:35:34 yes 2014-10-10 12:35:56 i think you'll have to implement it as a server, that sets up a unix socket listerner 2014-10-10 12:36:02 and write a client that connects to it 2014-10-10 12:52:53 i noticed a missing tcsh aports 2014-10-10 12:53:15 so I built one but tcsh coredumps. 2014-10-10 12:53:43 before I look into it, has someone already tried to aport tcsh? 2014-10-10 15:10:47 is there a berkeley-db in alpine ? 2014-10-10 15:11:01 db 2014-10-10 15:11:06 gotta go 2014-10-10 15:11:08 see u 2014-10-10 15:11:34 db-5.3.28-r0 << this one right? 2014-10-10 15:12:08 <[xming]> sounds right 2014-10-10 15:12:13 ok ty 2014-10-10 21:00:56 hey all I am planning on making a personal package mirror instance what is the approximate disk space do I need 2014-10-11 18:46:50 Watching netflix with html5 player - https://plus.google.com/104912707432334684242/posts/1gnbmo63xVa 2014-10-13 07:08:57 morning 2014-10-13 07:14:25 hi 2014-10-13 08:57:30 ncopa, you about? 2014-10-13 11:08:52 ncopa, when can i get excited about alpine-xen without python and perl? 2014-10-14 07:38:42 morning 2014-10-14 07:56:11 morning 2014-10-14 08:10:38 cd .. 2014-10-14 08:58:10 ncopa: do we have a definition for custom license? 2014-10-14 08:58:21 no 2014-10-14 08:58:40 i think the idea is that we ship the license in the package 2014-10-14 08:58:54 but i prefer avoid licence="custom" 2014-10-14 08:59:15 so what to provide? 2014-10-14 08:59:28 ? 2014-10-14 08:59:33 if i include it, whats the name 2014-10-14 08:59:58 LICENSE ? 2014-10-14 09:00:02 i think we normally put it in /usr/share/doc/licenses or similar 2014-10-14 09:00:19 i mean licence="custom" 2014-10-14 09:00:24 greetings mr ncopa :) 2014-10-14 09:00:50 seems we do: install -Dm644 COPYING "$pkgdir"/usr/share/licenses/$pkgname/COPYING 2014-10-14 09:00:54 hi xen_roger 2014-10-14 09:01:01 you want to prevent that. is there an alternative name? 2014-10-14 09:01:13 depends on what the license is? 2014-10-14 09:01:27 it doesnt specify, its custom i guess. 2014-10-14 09:01:49 for instance: https://github.com/discogs/discogs_client/blob/master/LICENSE 2014-10-14 09:02:21 that looks like bsd style license 2014-10-14 09:02:36 yes ... agree 2014-10-14 09:02:51 im not an expert on license :) 2014-10-14 09:03:23 https://fedoraproject.org/wiki/Licensing:BSD?rd=Licensing/BSD#2ClauseBSD 2014-10-14 09:03:49 yeah 2014-10-14 09:03:55 license="BSD" 2014-10-14 09:05:32 ncopa: seems interesting to have in our wiki 2014-10-14 09:05:51 what is the license on that information? can we copy it? ;-) 2014-10-14 09:08:00 :) 2014-10-14 10:01:01 ncopa: im excited about the next alpine-xen release. Have you removed xend python and perl from the iso? 2014-10-14 10:01:18 i think so yes 2014-10-14 10:01:38 those were pulled in as dependencies so they shoudl be removed from alpine-xen-3.1.x 2014-10-14 10:01:57 but i havent verified that it actually happens 2014-10-14 10:02:06 is that the next release? not 3.0.6? 2014-10-14 10:02:48 v3.0 is stable branch 2014-10-14 10:02:55 i dont want push this change to a stable branch 2014-10-14 10:03:14 might be users have done admin perl scripts 2014-10-14 10:03:16 if they upgrade 2014-10-14 10:03:23 and perl gets removed 2014-10-14 10:03:32 unexpected things might happen 2014-10-14 10:04:04 i dont think it will break things, but i dont want introduce unexpected surprises in a stable branch 2014-10-14 10:04:32 so v3.0 branch (next is v3.0.6) will not get this change 2014-10-14 10:04:43 and it will not get xen-4.4 2014-10-14 10:05:10 xen 4.4 (incl perl/python dep removal) will come with v3.1 2014-10-14 10:05:35 we should get 3.1.0_rc1 out soonish 2014-10-14 10:05:53 we need get 3.1.0_rc out soonish 2014-10-14 10:53:56 ncopa yeah, i take your point. perhaps 3.1 should come with a warning about perl and python removal :) 2014-10-14 11:00:07 there seems to be only one perl script in alpine-xen 2.7.9 /etc/ssl/misc/tsget 2014-10-14 11:00:53 oh and I'm ready to test 3.1 rc1 :) 2014-10-14 11:56:11 <[xming]> why can't ld find libuuid? 2014-10-14 11:56:19 <[xming]> /usr/lib/gcc/x86_64-alpine-linux-musl/4.8.2/../../../../x86_64-alpine-linux-musl/bin/ld: cannot find -luuid collect2: error: ld returned 1 exit status 2014-10-14 11:56:27 <[xming]> and I have libuuid installed 2014-10-14 11:57:20 <[xming]> ls /lib/libuuid.so* 2014-10-14 11:57:30 <[xming]> /lib/libuuid.so.1 /lib/libuuid.so.1.3.0 2014-10-14 11:57:39 isnt it part of util-linux? 2014-10-14 11:59:24 <[xming]> should be, the one I put in the deps was installing libuuid-2.24.2-r4 2014-10-14 11:59:41 did you add util-linux-dev? 2014-10-14 12:00:42 <[xming]> nope, but does it matter? 2014-10-14 12:01:11 you could try 2014-10-14 12:01:24 <[xming]> just tried 2014-10-14 12:03:29 <[xming]> this is weird 2014-10-14 12:05:48 <[xming]> does -dev pkg influence how ld works? 2014-10-14 12:06:00 <[xming]> I've even tried -L/lib 2014-10-14 12:06:33 <[xming]> clandmeter: thanks 2014-10-14 12:06:43 no, i dont think so 2014-10-14 12:07:54 <[xming]> I though dev was only .h and .a 2014-10-14 12:08:08 <[xming]> why linking to a .so would be affected? 2014-10-14 12:11:56 .so is symlinks 2014-10-14 12:12:40 you have /lib/libuuid.so.1 /lib/libuuid.so.1.3.0 2014-10-14 12:13:04 but how would ld know the exact filename if you did -luuid ? 2014-10-14 12:13:29 how does it know if it should use libuuid.so.1.3.0 or libuuid.s0.1.3.1? 2014-10-14 12:13:57 <[xming]> it isn't smart enough to follow the sym link? :D 2014-10-14 12:14:08 the .so symlink is there? 2014-10-14 12:14:13 <[xming]> yes 2014-10-14 12:14:15 <[xming]> ls /lib/libuuid.so* 2014-10-14 12:14:15 <[xming]> /lib/libuuid.so.1 /lib/libuuid.so.1.3.0 2014-10-14 12:14:17 <[xming]> no 2014-10-14 12:14:28 <[xming]> not the .so sorry 2014-10-14 12:14:34 thats what missing 2014-10-14 12:14:39 its shipped with -dev package 2014-10-14 12:14:48 and i thin its util-linux-dev as said 2014-10-14 12:15:14 <[xming]> it's still not there after I installed the -dev 2014-10-14 12:16:13 hi all, everything working out on 3.0.5 alpinexen :) 2014-10-14 12:16:28 <[xming]> Installed: Available: 2014-10-14 12:16:34 <[xming]> util-linux-dev-2.24.2-r4 = 2.24.2-r4 2014-10-14 12:16:37 ncopa-desktop:~/src$ apk info --who-owns /usr/lib/libuuid.so 2014-10-14 12:16:47 its there, just, now where you thought it would be 2014-10-14 12:16:52 just not* 2014-10-14 12:16:59 <[xming]> Oh 2014-10-14 12:17:02 <[xming]> thanks 2014-10-14 12:17:03 Frosh: nice! 2014-10-14 12:17:29 Frosh: i'd like to replace the gnu screen support with tmux 2014-10-14 12:17:37 possibly support both 2014-10-14 12:17:56 <[xming]> ncopa: it's basically the same question/issue I had with how abuild packages the -dev 2014-10-14 12:18:14 ncopa: I like screen, don't want to learn another thing 2014-10-14 12:18:20 :D 2014-10-14 12:18:31 ok i think that pretty much answered my question 2014-10-14 12:18:37 <[xming]> I like screen too, but it's heavy 2014-10-14 12:18:44 i prefer tmux 2014-10-14 12:18:53 <[xming]> like I like bash too 2014-10-14 12:19:02 i dont like bash :) 2014-10-14 12:19:30 ok, at least i know that i shouldnt just replace gnu screen with tmu 2014-10-14 12:19:31 tmux 2014-10-14 12:19:34 i'd need to support both 2014-10-14 12:20:42 <[xming]> I start using this about 5 months ago http://www.brain-dump.org/projects/dvtm/ 2014-10-14 12:20:52 <[xming]> in screen/tmux 2014-10-14 12:22:11 <[xming]> ncopa: you lua bot works very nice, got me started to look into lua 2014-10-14 12:22:37 I can confirm this bug too: https://bugs.alpinelinux.org/issues/3107 lvm still works 2014-10-14 12:22:53 <[xming]> BTW I've add http://alpinelinux.org/ in my yacy index 2014-10-14 12:23:11 <[xming]> people searching on yacy will find alpine 2014-10-14 12:23:19 <[xming]> should I add the wiki too? 2014-10-14 12:23:35 lvm works, it just gives a warning that it could not protect lvm from beeing swapped out 2014-10-14 12:25:08 ScrumpyJack: you can try xen-4.4 here: http://dev.alpinelinux.org/~ncopa/xen/ 2014-10-14 12:25:13 What's strange is I'm not sure where this other issue when I startup centos7 from time to time, it freezes on bootingup, it will be in blocked state and Time would be 1.9 2014-10-14 12:28:43 I'm not sure how to troubleshoot from there :( 2014-10-14 12:37:27 <[xming]> http://search.yacy.net/yacysearch.html?query=alpine+linux&verify=ifexist&contentdom=text&nav=filetype%2Cprotocol%2Chosts%2Cauthors%2Cnamespace%2Ctopics&startRecord=0&indexof=off&meanCount=5&resource=global&prefermaskfilter=&maximumRecords=10 2014-10-14 12:42:15 <[xming]> ACTION lost his tail 2014-10-14 13:13:12 <[xming]> hmm I have a problem with GNU make 4.0 in the 3.0.5, upgrading to 4.1 in edge solves it 2014-10-14 13:13:40 <[xming]> make 4.0 uses over 4GB of RAM and was still increasing 2014-10-14 13:14:08 <[xming]> for a small dir in doc/ 2014-10-14 13:15:19 ncopa: is xen_roger1 the xen specialist? 2014-10-14 13:19:03 i use screen and tmux all the time. is there a problem with keeping both? 2014-10-14 13:20:46 ncopa: thanks. how is alpine-xen so big? :) its still 118.9M without perl and python? 2014-10-14 13:28:41 ScrumpyJack: built new iso 2014-10-14 13:28:46 it had xen-xend 2014-10-14 13:28:51 which is no longer supported 2014-10-14 13:31:13 118.9M is not big at all ;) 2014-10-14 13:31:30 alpine-xen-141014-x86_64.iso 2014-Oct-14 13:28:18 109.0M 2014-10-14 13:36:57 alpine-dev-edge:~$ apk info -s python perl 2014-10-14 13:37:02 python-2.7.8-r0 installed size: 2014-10-14 13:37:03 37572608 2014-10-14 13:37:03 45010944 2014-10-14 13:37:03 perl-5.20.1-r0 installed size: 2014-10-14 13:37:26 that's 82 megs no? 2014-10-14 13:38:00 alpine-xen was 125 megs, so i thought it would be much smaller without python and perl :) 2014-10-14 13:38:35 ScrumpyJack: why should it be smaller than that? is there any reason ? 2014-10-14 13:39:01 xen01:~# ls -l /media/usb/apks/x86_64/p[ey]* 2014-10-14 13:39:01 -rwxr-xr-x 1 root root 11328798 Oct 8 12:54 /media/usb/apks/x86_64/perl-5.18.2-r0.apk 2014-10-14 13:39:02 -rwxr-xr-x 1 root root 10005546 Oct 8 12:54 /media/usb/apks/x86_64/python-2.7.8-r0.apk 2014-10-14 13:39:22 I added vlan bridge and some other pkg to my iso i missed ... is arround 200M now 2014-10-14 13:39:55 when openvswitch is stable on alpine would be nice if it is included in xen image by default ... 2014-10-14 13:40:04 ScrumpyJack: the .apk are gz compressed 2014-10-14 13:40:11 so the diff should be ~21MB 2014-10-14 13:41:06 the runtime tmpfs should be 82MB smaller though 2014-10-14 14:48:33 ncopa, can we figure out if phpmyadmin would work with php v5.6x ? 2014-10-14 14:48:34 http://sourceforge.net/p/phpmyadmin/mailman/message/32284478/ 2014-10-14 14:49:54 maybe ask upstream? 2014-10-14 14:50:07 ok 2014-10-14 14:50:17 also could you pls see http://bugs.alpinelinux.org/issues/2683 2014-10-14 14:52:53 isnt it shipped witih php-5.6? 2014-10-14 14:53:15 but don't know if needs a config option to enable it build 2014-10-14 14:53:30 would check 2014-10-14 14:53:30 can you check if php-5.6 built in edge has it? 2014-10-14 14:54:17 ncopa: duh! silly me 2014-10-14 14:55:19 hmm php-dbg-5.6.1-r0.apk is there ? 2014-10-14 14:58:14 but than it is also in php-dbg-5.5.16-r0.apk v3.0 repo 2014-10-14 14:58:34 so guess its not phpdbg 2014-10-14 15:36:28 there is --enable-phpdbg option 2014-10-14 20:57:48 cy 2014-10-15 07:34:35 configure: error: --with-libxul-sdk is not supported anymore. 2014-10-15 07:34:41 we cannot do shared xulrunner anymore 2014-10-15 07:34:47 firefox is annoying me 2014-10-15 07:35:19 i wonder what this means for openjdk7 2014-10-15 07:35:54 do we need build xulrunner separate or can we have openjdk7 to build with firefox-dev? 2014-10-15 08:04:58 09:34 <@ncopa> we cannot do shared xulrunner anymore 2014-10-15 08:05:01 what do you mean? 2014-10-15 08:05:41 ncopa> configure: error: --with-libxul-sdk is not supported anymore. 2014-10-15 08:05:57 firefox 33 does nosupport shared xulrunner 2014-10-15 08:06:20 I see 2014-10-15 08:06:34 I don't think that Arch builds openjdk with Fx or xulrunner 2014-10-15 08:06:59 ok 2014-10-15 08:07:07 and icedtea-web? 2014-10-15 08:07:17 java plugin for firefox? 2014-10-15 08:07:41 it is here, and worked fine some time ago 2014-10-15 08:07:58 dont need xulrunner for that? 2014-10-15 08:08:18 maybe i can do a firefox-dev subpackage 2014-10-15 08:09:04 the idea with separate xulrunner was to get thunderbird compiled 2014-10-15 08:09:20 but i kinda gave that up long time ago 2014-10-15 11:30:52 <[xming]> how can I make abuild -r to prefer my local (newer) pkg for deps instead of pulling older ones from the remote repo? 2014-10-15 11:31:45 add repo to /etc/apk/repositaries where your package is 2014-10-15 11:32:53 <[xming]> I think I've done done 2014-10-15 11:33:03 if package's release is higher then in main repo you can istall it 2014-10-15 11:35:16 <[xming]> it's in /etc/apk/repositories 2014-10-15 11:35:24 <[xming]> the issues is that APKINDEX.tar.gz isn't updated 2014-10-15 11:35:50 <[xming]> pkgs I've built ealier are in there, but not the one I am trying to install 2014-10-15 11:36:46 <[xming]> what could be wrong? and how can I regenerte APKINDEX? 2014-10-15 11:37:07 apk update ? 2014-10-15 11:37:59 abuild update_abuildrepo_index 2014-10-15 11:38:15 in next version of abuild its gonna be: abuild index (in current git) 2014-10-15 11:39:02 <[xming]> thanks, ncopa any idea why this is happening? IIRC this happened a few times, and before I just deleted the whole packages/ and rebuilt everything 2014-10-15 11:39:37 <[xming]> ah it's updated 2014-10-15 11:40:11 <[xming]> libzdb version bump http://sprunge.us/fWZd 2014-10-15 12:02:06 <[xming]> ncopa: I am trying to build http://seafile.com/en/home/ but have to build all its dependencies, are you willing commit this (5~6 pkgs in total)? 2014-10-15 12:04:52 probably yes 2014-10-15 12:05:07 depends on how good they are 2014-10-15 12:05:24 if they need to much fixing i'll probably need to do it later 2014-10-15 12:05:37 <[xming]> I am not ready yet 2014-10-15 12:05:43 ok 2014-10-15 12:05:50 im fighting firefox-33 2014-10-15 12:06:02 <[xming]> I am running it here abeit under Gentoo 2014-10-15 12:06:11 <[xming]> what's wrong? 2014-10-15 12:06:20 repackagin 2014-10-15 12:06:32 we used to ship separate xulrunner package 2014-10-15 12:06:44 <[xming]> ah 2014-10-15 12:06:46 now they dont support firefox with separate xulrunner 2014-10-15 12:06:50 so i have to merge it 2014-10-15 12:06:56 and clean up subpackages etc 2014-10-15 12:07:03 every rebuild takes ages... 2014-10-15 12:07:07 <[xming]> to me it builds the same way as 32 (bundled xulrunner) 2014-10-15 12:07:34 <[xming]> ccache? 2014-10-15 12:07:49 its the unbundle xulrunner that takes time 2014-10-15 12:07:52 i have ccache 2014-10-15 12:08:20 and i have 12 cores + hyperthreading (24 'cpus') 2014-10-15 12:08:23 <[xming]> so apine is still trying to use seperate xulrunner? 2014-10-15 12:08:31 its still slow 2014-10-15 12:08:34 nah 2014-10-15 12:08:39 i just gave that up 2014-10-15 12:09:00 <[xming]> but you are trying to make xulrunner as a subpkg from the main firefox? 2014-10-15 12:09:07 dunno yet 2014-10-15 12:09:34 <[xming]> FF takes like 45min here (4 cores :( 2014-10-15 12:09:42 yes its slow 2014-10-15 12:10:06 <[xming]> 12 cores is like 15 min :p 2014-10-15 12:11:21 <[xming]> aren't there still some -j1 in the firefox build? 2014-10-15 12:15:24 <[xming]> my local APKINDEX.tar.gz is really not getting updated, happened again 2014-10-15 13:03:41 ncopa, thanks for incron, nilfs-utils 2014-10-15 13:07:17 and php 5.6x has --enable-phpdbg option 2014-10-15 13:08:45 if separate pkg is created than could be named php-phpdbg-5.6.1-r0.apk 2014-10-15 13:17:48 firefox build time: 2014-10-15 13:17:51 >>> firefox: Signing the index... 2014-10-15 13:17:51 real 24m 47.20s 2014-10-15 13:17:51 user 4h 4m 41s 2014-10-15 13:17:51 sys 14m 39.57s 2014-10-15 13:17:58 too much... 2014-10-15 13:18:26 whatta 2014-10-15 13:18:32 >>> firefox*: Package size: 135.4 MB 2014-10-15 13:30:39 vkrishn: add the phpdbg stuff to #2683 so i dont forget it 2014-10-15 13:30:40 thanks 2014-10-15 13:31:52 ok, but have not tested if it would build on AL 2014-10-15 13:35:09 done 2014-10-15 13:39:13 thanks 2014-10-15 14:26:23 new openssl 2014-10-15 14:26:27 i need to go 2014-10-15 14:26:35 perhaps someone can bump it if there's no conflicts in patches 2014-10-15 14:26:39 otherwise i'll do it tm. 2014-10-15 14:27:33 no conflicts apparently 2014-10-15 14:27:43 i'll bump it 2014-10-15 14:30:29 ncopa: do you know if alpinexen will support pygrub2 ? 2014-10-15 14:31:35 he... Frosh is back ;) 2014-10-15 14:31:56 did you get you centos running? 2014-10-15 14:32:06 Frosh: i dont know. is it built by default xen-4.4? 2014-10-15 14:33:19 pygrub was moved to subpackage py-xen 2014-10-15 14:33:39 i dunno fif xen-4.4. ships pygrub2 or if its old version 2014-10-15 14:36:38 xen_roger: nope, something with centos i think with mkfs.ext2 on some uuid thats not mounted on time? 2014-10-15 14:37:18 the other thing which is less of a issue, is i have to xl unpause a domain sometimes 2014-10-15 14:37:31 Frosh: strange ... got it runnging yesterday ... also with pvgrub. 2014-10-15 14:39:37 Frosh: sounds like you have an general issue with your setup 2014-10-15 14:40:23 http://wiki.xen.org/wiki/Xen_4.4_Feature_List#GRUB_2_Support_of_Xen_Project_PV_Images_.28External.29 2014-10-15 15:02:41 ncopa, how does mod-sflow work ? 2014-10-15 15:03:17 vkrishn: have no idea 2014-10-15 15:03:35 does it not require Host-sFlow daemon ? 2014-10-15 15:04:23 I think adding nginx-sflow-module would also be nice 2014-10-15 15:06:14 nice initiative http://www.linuxfoundation.org/news-media/announcements/2014/09/telecom-industry-and-vendors-unite-build-common-open-platform 2014-10-15 15:34:53 ncopa: 25 minutes doesn't sound so bad 2014-10-15 20:54:20 that was fast 2014-10-15 20:54:25 only 5 hours 2014-10-16 08:25:09 ncopa, pushing new openjdk soon. seems that it compiles just fine. 2014-10-16 08:25:24 i'm also planning to cherry-pick apk-tools and musl to 3.0-stable. any objections? 2014-10-16 09:19:26 ncopa: we use syslog-ng successfully on our installations, so could you please move it from testing to main? 2014-10-16 09:19:54 you can setup myself as maintainer if neccessary, no problem for me 2014-10-16 09:38:18 StarWarsFan: why is there a post-upgrade script that does_ 2014-10-16 09:38:23 rc-service syslog-ng start 2014-10-16 09:38:49 we should not start/stop anything from installs cripts 2014-10-16 09:39:08 example: apk upgrade --root /some/chroot 2014-10-16 09:39:31 is you happen to have syslog-ng installed there, and it gets upgraded, then it will also start within the chroot 2014-10-16 09:39:38 that is probably not what you want 2014-10-16 09:40:32 it also tries stop syslog-ng and removes it from runlevel boot on deinstall 2014-10-16 09:40:40 i dont think we want do that either 2014-10-16 09:41:06 rm -f /etc/syslog-ng/syslog-ng-*.std 2014-10-16 09:41:07 rm -f /etc/syslog-ng/*.apk* 2014-10-16 09:41:10 thats preinstall 2014-10-16 09:41:56 if you copy some package into /etc/syslog-ng/ it will be removed when you do apk add syslog-ng 2014-10-16 09:43:15 i think we can just delete all the install scripts 2014-10-16 09:45:19 i c, will check this with my colleague 2014-10-16 09:45:26 thx for the input 2014-10-16 09:48:36 basically, we try not autostart/autofix anything from install scripts 2014-10-16 09:48:56 i mean start/stop services 2014-10-16 09:49:14 nor do we do anything interactive (ask uwer for confirmation) 2014-10-16 09:49:17 user* 2014-10-16 09:49:49 and we try avoid fix permissions or create/delete files/dirs 2014-10-16 10:48:23 <[xming]> minetest abuild -> http://sprunge.us/bGjL 2014-10-16 10:50:00 <[xming]> minetest-server abuild -> http://sprunge.us/EgTM 2014-10-16 10:51:27 <[xming]> minetest-server has been tested for a few days w/o issues, but it depends on minetest-common which is provided by minetest (client), which I can't test because I don't have Alpine with X11 2014-10-16 10:54:42 ncopa: ok, i think we create a new patch and update the current state of the package 2014-10-16 10:55:02 there are some changes on our own version which are not in the alpine-repo at mo... ;-) 2014-10-16 13:00:47 afternoon 2014-10-16 13:55:19 Hello. 2014-10-16 13:55:39 I noticed that bind's default config enables recursion for localhost. 2014-10-16 13:56:21 This seems rather odd, since it means that an authoritative nameserver is liable to serve records for which it is not authoritative when queried from localhost. 2014-10-16 13:58:04 Since running the same daemon as authoritative nameserver and recursive resolver is not especially advised nowadays, perhaps the default config should be modified to disable this? 2014-10-16 14:00:55 sound good to me 2014-10-16 14:01:30 hl: do you have a suggestion to a new default config? 2014-10-16 14:01:47 ncopa: yes, inside named.conf, options{}; block, add: 2014-10-16 14:01:58 allow-recursion { none; }; 2014-10-16 14:02:02 recursion no; 2014-10-16 14:03:56 woudl allow-recursion { 127.0.0.1/32; }; make sense? 2014-10-16 14:04:28 I think that's what bind defaults to, but I don't think it makes a great deal of sense. Recursion simply shouldn't be done in the same daemon. 2014-10-16 14:04:30 http://cr.yp.to/djbdns/separation.html 2014-10-16 14:05:33 true 2014-10-16 14:05:34 but 2014-10-16 14:05:37 the config also has: 2014-10-16 14:05:43 zone "." IN { 2014-10-16 14:05:43 type hint; 2014-10-16 14:05:43 file "named.ca"; 2014-10-16 14:05:43 }; 2014-10-16 14:05:53 which i suppose does not make sense without recursion 2014-10-16 14:06:28 Yes, pretty much. Perhaps you should provide two example files, named.conf.authoritative and named.conf.recursive, and make them rename one? 2014-10-16 14:07:09 Actually, gimme a sec, I'll put some commented files together and send a patch. 2014-10-16 14:08:04 would be great 2014-10-16 14:08:06 thanks! 2014-10-16 14:29:45 ncopa: hello, about grantlee and qimageblitz (new aports), should i resend patch, because of wrong dependencie ? 2014-10-16 14:30:20 i just havent had time to look at them 2014-10-16 14:30:25 isn't ... 2014-10-16 14:30:28 there was also a bunch of kde patches? 2014-10-16 14:30:35 Recursion no 2014-10-16 14:30:45 Allow-transfer { none;}; 2014-10-16 14:30:47 better? 2014-10-16 14:30:51 yes, i already have kde-core meta package also, but it will be huge patch 2014-10-16 14:31:24 maybe have a git repo on github so i can git pull them all in one shot? 2014-10-16 14:31:35 if u want i can resend wrong patches 2014-10-16 14:32:03 sure ;), http://github.com/k0r10n/aports 2014-10-16 14:32:47 um 2014-10-16 14:32:55 i doubt i can pull that 2014-10-16 14:33:14 ok 2014-10-16 14:33:14 it does not share same origin 2014-10-16 14:33:18 yes 2014-10-16 14:33:58 clone this: https://github.com/alpinelinux/aports 2014-10-16 14:34:10 and push your stuff there 2014-10-16 14:34:16 to your clone 2014-10-16 14:34:20 then i can pull it 2014-10-16 14:34:34 ok, ty 2014-10-16 14:45:14 i thought allow-recursion { none; }; is only relevant if you have recursion yes 2014-10-16 14:45:32 k0r10n: please prefix subject in patches with which package it is about 2014-10-16 14:45:33 Pretty much, but it can't hurt. 2014-10-16 14:45:38 0c0410c6d0159dc7e7baf1ac30e15dd46d7e06bd Added -fPIC flag to fix pci_malloc relocation 2014-10-16 14:46:56 k0r10n: image a release notes with this: http://sprunge.us/iATS 2014-10-16 14:47:30 it is not really clear what your patch actually fixes 2014-10-16 14:48:33 this is clearer: http://sprunge.us/WYdX 2014-10-16 14:51:58 hl: what the default for Allow-transfer? 2014-10-16 14:52:18 hl: isn't it { any;}; ? 2014-10-16 14:52:48 hl: perhaps that should be { none;}; if you are tidying up :) 2014-10-16 14:52:57 ncopa: the problem with pciutils is in libpci.a, when kde-runtime tried to link it on 32-bit i have TEXTREL, and on 64 only error, so i don't know what 2014-10-16 14:53:04 's the best way to do it 2014-10-16 14:53:05 :) 2014-10-16 14:53:26 how to explain it in one string :) 2014-10-16 14:54:21 k0r10n: the point is that you include which package it is 2014-10-16 14:54:29 ncopa: yes, understood 2014-10-16 14:54:39 sry :) 2014-10-16 14:54:43 np 2014-10-16 14:54:54 you dont think about it til you see the release notes... 2014-10-16 14:54:59 or the git log 2014-10-16 14:55:09 i thought you mean about explanation...... 2014-10-16 14:55:38 i'll reply it to the mailing list so others can learn too 2014-10-16 14:55:46 you are not the first... 2014-10-16 14:57:09 hm 2014-10-16 14:57:13 libpci.a 2014-10-16 14:57:21 i think we should remove it 2014-10-16 14:57:47 and use the dynamic lib instead 2014-10-16 14:57:56 hm 2014-10-16 14:58:01 there is no dynamic libpci 2014-10-16 15:52:50 patch sent. 2014-10-16 21:57:05 hi 2014-10-16 21:57:43 I was curious about the fact that both android and alpine use musl 2014-10-16 21:58:19 So is there a chance that this can lead to some kind of interoperability 2014-10-16 21:58:27 Or better, portability? 2014-10-16 22:01:00 any idea 2014-10-16 22:01:01 ? 2014-10-16 22:29:04 :-) 2014-10-16 22:29:10 hi 2014-10-16 22:29:42 Did Android ditch bionic for musl already? You can always put an Alpine chroot on a phone :) 2014-10-16 22:30:08 yeah, Android uses musl 2014-10-16 22:30:22 and I was reading about wayland tricks 2014-10-16 22:30:41 that Canonical introduced to use graphical libraries 2014-10-16 22:30:49 and drivers 2014-10-16 22:30:56 only available to Android 2014-10-16 22:30:59 many vendors 2014-10-16 22:31:18 don't release properly-licensed drivers for Linux 2014-10-16 22:31:31 but they often provide drivers for Android 2014-10-16 22:31:47 binary or sometimes sources 2014-10-16 22:32:43 Yeah, proper graphics acceleration on Linux/ARM sounds like more of a pain than it is on the desktop. 2014-10-16 22:34:55 Hrmm, I should blow away my OpenWRT chroot on my Android and replace it with Alpine's new ARM port. 2014-10-16 22:35:54 could you please explain 2014-10-16 22:36:02 how to get an Alpine chroot on a phone? 2014-10-16 22:36:12 may be create a tree 2014-10-16 22:36:16 and then chrooting into? 2014-10-16 22:37:30 You have to have your phone rooted, then drop a statically linked busybox somewhere writeable/executable and then go to town. 2014-10-16 22:39:20 I hacked on the initrd to integrate OpenWRT into Android's init. 2014-10-16 22:39:42 and that worked? 2014-10-16 22:39:47 Eventually 2014-10-16 22:39:49 do you use that on your router? 2014-10-16 22:41:28 Nah, at meetings, I'm a bit familiar with OpenWRT because of the routers though. This was before I really looked into Alpine. 2014-10-16 22:43:45 Alpine seems good to me 2014-10-16 22:44:01 It's fine, with a few glitches 2014-10-16 22:44:12 and some shortcomings that I want to fix 2014-10-16 22:45:47 But I don't know how far it can go 2014-10-16 22:46:02 Yeah, I'm using it on resource constrained virtual servers. Anything in particular you want to fix? 2014-10-16 22:47:19 I'm using it with XFCE 2014-10-16 22:47:45 and I want to make it more simple to use 2014-10-16 22:48:08 It seems that gnome uses a volume manager 2014-10-16 22:48:30 to list removable devices 2014-10-16 22:48:47 there isn't such thing in Alpine 2014-10-16 22:48:49 isn't it? 2014-10-16 22:49:29 or maybe it has 2014-10-16 22:49:34 because it has Gnome 2014-10-16 22:51:06 I'm not really familiar with Alpine on the desktop. I get the impression that integration between the big desktop environments and non-systemd/non-udev is going to get worse. 2014-10-16 22:53:07 I've read about systemd/udev 2014-10-16 22:53:23 But I didn't worked with them much 2014-10-16 22:54:08 The whole UNIX/Linux environment is changing 2014-10-16 22:54:21 The fall of X and emergence of Wayland 2014-10-16 22:54:41 and that Systemd with several stuff attached to it 2014-10-16 22:55:04 and that Android thing 2014-10-16 22:55:28 which partitioned the Java community 2014-10-16 22:59:12 I've got mixed feelings on systemd. On one hand, I really like a bunch of the features they offer, but the politics of it are too overwhelming. 2014-10-16 22:59:51 One thing I read in the developers blog 2014-10-16 23:00:10 was the ability to boot without a shell 2014-10-16 23:00:30 if remember well 2014-10-16 23:00:45 and these kind of things 2014-10-16 23:00:50 that speed up boot process 2014-10-16 23:01:01 Alpine is fast 2014-10-16 23:01:08 but the boot process is actually slow 2014-10-16 23:01:55 I'm eager to see the impact of systemd 2014-10-16 23:02:40 on boot performance 2014-10-16 23:03:24 this is from 2009 2014-10-16 23:03:25 lwn.net/Articles/299483/ 2014-10-16 23:03:33 I especially like the "socket activation" they've implemented. Essentially systemd will open up a UNIX/TCP/UDP/inotify/pty/fifo/netlink/etc... file descriptor for you (the daemon) so it'll be ready for you once you start up with potentially queued up clients. 2014-10-16 23:04:40 it makes the startup dependencies implicit, and you can even lazy load services using file descriptor readiness calls like poll()/select()/epoll() 2014-10-16 23:04:52 I'm curious to see how far we can go with systemd and boot optimization. 2014-10-16 23:07:47 It seems that Linus is OK with systemd 2014-10-16 23:08:43 Because one that thinks about his driver/application/etc portability on different Unixes has a very different mindset with who creates platforms 2014-10-16 23:08:49 systemd is plenty fast and I'm on an HDD. 2014-10-16 23:09:20 I should give it a try 2014-10-16 23:10:27 Doesn't grsec break systemd? 2014-10-16 23:11:24 I'm actually running Arch on my desktop/laptop. 2014-10-16 23:12:10 I'm happy with Debian/Ubuntu 2014-10-16 23:12:23 So I've stopped changing my distros on my desktop/laptop 2014-10-16 23:12:50 But I'm using Alpine for those devices with less CPU power 2014-10-16 23:13:14 Is there a specific thing that Arch is great in it? 2014-10-16 23:13:36 I'm sure there are things 2014-10-16 23:14:34 Let me ask what I forgot 2014-10-16 23:14:46 I eneded my distro hopping ways 8 years ago with Arch. I like to think of it as the midpoint between Slackware and Gentoo. 2014-10-16 23:15:22 Do you know which Alpine package I should install to have gnome volume management; I mean automatic mounting of removable storage? 2014-10-16 23:15:31 Gentoo? 2014-10-16 23:15:57 That needs a lot of time! 2014-10-16 23:16:16 How long it takes to build a system from scratch? 2014-10-16 23:16:28 or at least the minimal system it offers? 2014-10-16 23:17:08 I'm installing nautilus 2014-10-16 23:17:13 Hope that it works 2014-10-16 23:17:14 :-) 2014-10-16 23:18:13 Can't help you with the volume manager. Gentoo took forever when I last used it, but I was on some pretty slow hardware at the time though. 2014-10-16 23:19:31 I wanted to try gentoo 2014-10-16 23:19:45 But I read several things about it 2014-10-16 23:19:50 and I came to a conclusion 2014-10-16 23:20:06 that compiling everything by yourself 2014-10-16 23:20:33 can't give you too much speedup 2014-10-16 23:20:54 switching from libc to musl 2014-10-16 23:21:14 seems to make visible difference 2014-10-16 23:21:35 and obvious space reduction 2014-10-16 23:22:08 Once I thought about compiling everything with Intel compiler 2014-10-16 23:22:13 but that's a real headache! 2014-10-16 23:22:33 even compiling kernel itself with ICC is a real headache 2014-10-16 23:23:59 Did it end up working? 2014-10-16 23:24:18 xubuntu does not have nautilus 2014-10-16 23:24:19 but it has the ability to mount removable devices 2014-10-16 23:24:41 I don't know where that functionality lies 2014-10-16 23:25:11 Ubuntu is great because it works out of the box 2014-10-16 23:25:21 you don't have to spend hours and maybe days 2014-10-16 23:25:25 to configure your desktop 2014-10-16 23:28:59 I gave up searching and I just tried to install nautilus right now 2014-10-16 23:29:13 That's if you agree with the developers' decisions :) 2014-10-16 23:30:29 The bad thing here 2014-10-16 23:30:45 is that you aren't asked about what you are forced to install 2014-10-16 23:31:02 It was a question for me that why Debian/Ubuntu asks some time 2014-10-16 23:31:09 if it installs a package or not 2014-10-16 23:31:19 when I used "apt-get install" 2014-10-16 23:31:32 and at last I found that 2014-10-16 23:31:48 whenever it needs to install only the requested package 2014-10-16 23:31:53 it doesn't ask 2014-10-16 23:31:58 because you ordered to do so 2014-10-16 23:32:12 but when it needs to install other dependencies, for sure it asks 2014-10-16 23:32:33 because it needs to confirm the installation of that additional packages 2014-10-16 23:32:41 I think it would be great 2014-10-16 23:32:59 to have such confirmation in APK 2014-10-16 23:36:01 installing nautilus didn't help 2014-10-16 23:36:25 I'll try to log in GNOME session 2014-10-16 23:36:37 but I think its GVFS or something like that 2014-10-16 23:40:31 I'll be back 2014-10-16 23:43:04 Good luck with automounting, unfortunately I'm the wrong person to ask. 2014-10-16 23:45:36 installing gvfs-* 2014-10-16 23:45:38 :-/ 2014-10-16 23:54:31 I came across this bug 2014-10-16 23:54:35 https://bugzilla.xfce.org/show_bug.cgi?id=9193 2014-10-17 00:47:38 it seems that I don't have HAL 2014-10-17 00:48:00 and I should have 2014-10-17 00:48:06 http://goodies.xfce.org/projects/thunar-plugins/thunar-volman 2014-10-17 00:52:32 is it hard to build such a thing for Alpine? 2014-10-17 00:56:24 oh 2014-10-17 00:56:26 it says 2014-10-17 00:56:27 HAL is now deprecated on most GNU/Linux distributions, such as parts of Ubuntu with functionality being merged into udev as of 2008–2010. Previously, HAL was built on top of udev. 2014-10-17 00:56:42 maybe it's ok in XFCE 4.12? 2014-10-17 01:15:15 ok thanx 2014-10-17 01:15:16 bye 2014-10-17 03:36:26 Ok 2014-10-17 03:36:34 Things get better 2014-10-17 03:36:50 and now I know that udev+udisks+udisks2+gvfs is needed 2014-10-17 03:37:06 to have removable disks automatically mounted in XFCE 2014-10-17 03:38:02 it is good to mention it in http://wiki.alpinelinux.org/wiki/XFCE_Setup 2014-10-17 03:38:36 It is also needed that dbus is added to startup daemons 2014-10-17 03:38:43 rc-update add dbus 2014-10-17 03:38:52 it worked for me 2014-10-17 03:39:23 with a little problem that it uses ntfs rather than ntfs-3g 2014-10-17 03:39:41 and ntfs is not available in the kernel shipped with Alpine 2014-10-17 03:40:01 so it is needed to modify udisks2 configurations 2014-10-17 03:40:04 I think 2014-10-17 03:43:08 That's it 2014-10-17 03:43:10 bye 2014-10-17 06:22:51 ncopa: this patch is not enough 2014-10-17 06:22:55 https://github.com/shimmerproject/Greybird/pull/70 is also needed 2014-10-17 06:23:06 morning 2014-10-17 06:23:09 i just saw that 2014-10-17 06:23:10 and theme still shows white background here and there 2014-10-17 06:23:13 thanks 2014-10-17 06:23:15 gnome sucks 2014-10-17 06:23:22 :) 2014-10-17 06:23:31 yeah, its sad what has become of it 2014-10-17 06:23:38 its sad what it has become 2014-10-17 06:24:37 barthalion: do you have patches for fixing those whit background here and there? 2014-10-17 06:24:45 I wish I have 2014-10-17 06:24:51 i see that tabs in firefox looks bad 2014-10-17 06:24:51 haven't found anything yet 2014-10-17 06:30:35 hm 2014-10-17 06:30:54 the spinner does not look good with greybird 2014-10-17 07:14:29 <[xming]> morning, I have aports for minetest and its server http://sprunge.us/HFOj and http://sprunge.us/ETjd 2014-10-17 07:16:13 <[xming]> ncopa: if you have some time can you please review comment above? It's my first aport with subpkg 2014-10-17 07:16:33 why the separated APKBUILDs? 2014-10-17 07:16:40 you can run cmake twice 2014-10-17 07:17:24 <[xming]> ah right! 2014-10-17 07:17:47 the rest looks ok 2014-10-17 07:17:59 yeah better run it twice 2014-10-17 07:18:06 from same apkbuild 2014-10-17 07:18:06 so if ncopa is fine with being set as a maintainer, I'd push it 2014-10-17 07:18:49 make the -server a subpackage of minetest 2014-10-17 07:19:10 you could have separate build dir too 2014-10-17 07:20:03 <[xming]> I can run camke ../src/foobar 2014-10-17 07:20:14 <[xming]> is there any aport func to facilitate that? 2014-10-17 07:20:27 yes or: mkdir build-server && cd build-server && cmake .. 2014-10-17 07:20:46 no special func for it 2014-10-17 07:21:52 <[xming]> okay more fiddling 2014-10-17 07:25:45 <[xming]> BTW something else, I really like alpine, might even switch some bare metals to alpine, but I want more the build from the source distro (with pkgs of my own CFLAGS), is there a way to rebuild all my world + its dependencies and install them over the pkgs from the repo? 2014-10-17 07:27:53 [xming]: you could clone the aports and build your own repo with special CFLAGS,but maybe not what you want?! 2014-10-17 07:29:56 hi guys, i'm looking #3443 2014-10-17 07:30:12 Got this error: 2014-10-17 07:30:12 poll-util.h:51:17: error: unknown type name 'nfds_t' 2014-10-17 07:30:52 this is the content of poll-util.h http://sprunge.us/WQfW 2014-10-17 07:31:30 Any hint? 2014-10-17 07:31:39 <[xming]> xen_roger1: yes, but that way I have to build the whole repo :/ 2014-10-17 07:33:07 just have your repo above ours 2014-10-17 07:33:23 [xming]: with some scripting you could handle it ;) - yes I get your point ... always have to recompile some packages like e.g. squid. 2014-10-17 07:34:32 <[xming]> I was wondering if alpine has already some thing like that to rebuild the world 2014-10-17 07:35:11 <[xming]> so I need a boostrapping env 2014-10-17 07:37:09 <[xming]> I like alpine of its basic and DIY, but sometimes I am too lazy and just want things done :p 2014-10-17 07:37:26 somthing like doing a ssh from build to target get list of pkg ( apk info | sort ) and then loop over them an rebuild with our flavour into local repo wich you could add as barthalion said above of the others... not nice but should work 2014-10-17 07:38:36 <[xming]> and some git post-commit hooks to build updates automatically :D 2014-10-17 07:39:25 fcolista2: seems like poll.h should provide nfds_t 2014-10-17 07:39:40 http://pubs.opengroup.org/onlinepubs/009695399/basedefs/poll.h.html 2014-10-17 07:40:30 [xming]: what kindof changes are you going to set? only CFLAGS? 2014-10-17 07:41:09 <[xming]> I thought with a version update (3.0.4 -> 3.0.5) that edge moves to 3.0.x but that isn't the case 2014-10-17 07:41:31 ncopa: ye 2014-10-17 07:41:32 yes 2014-10-17 07:41:42 and in /usr/include/poll.h exists 2014-10-17 07:42:02 seems like poll-util.h is not able to read/parse/look at poll.h 2014-10-17 07:42:17 <[xming]> xen_roger1: CFLAGS, LDFLAGS, just tinking when I feel like, maybe locally some weird/experiental patches/versions 2014-10-17 07:46:31 [xming]: I have some ideas to manage that ... ugly but should work.. maybe I check this out next days ... atm I build the packages one by one ... would be nice to have this done automaticly. 2014-10-17 07:50:25 <[xming]> and I was hoping that sudo apk -f add would reinstall the already istalled package 2014-10-17 08:00:38 <[xming]> minetest aport: http://sprunge.us/KRUf 2014-10-17 09:24:45 This is how abuild -r ends about yadifa 2014-10-17 09:25:29 this is the apkbuild http://sprunge.us/TUDO 2014-10-17 09:25:43 this is a patch to disable execinfo.h http://sprunge.us/BCRA 2014-10-17 09:26:15 If someone has a clue about why the poll error occurs, ping me 2014-10-17 10:00:00 <[xming]> fcolista2: moved #include out of #ifndef _POLL_H works 2014-10-17 10:00:54 <[xming]> something makes it thinks it's defined, but a grep shows nothing 2014-10-17 10:06:51 it will result in a linker error 2014-10-17 10:09:12 [xming] thanks. It worked. Now there is another issue, but i'll back later if i need help. 2014-10-17 10:09:15 thx 2014-10-17 10:10:14 <[xming]> yeah I get a ld error, pointing to backstrace and that is in execinfo? 2014-10-17 10:10:40 yeah 2014-10-17 10:11:06 right 2014-10-17 10:11:14 this is because of execinfo.h 2014-10-17 10:11:40 i think that debug shoudl be optional 2014-10-17 10:12:34 i'm not a c coder, but something like #ifdef DEBUG #include #endif would be better. 2014-10-17 10:15:04 <[xming]> fbsd has it http://www.freshports.org/dns/yadifa/ 2014-10-17 10:17:55 yes but they dont have patches. i guess it checks for fbsd. 2014-10-17 10:19:40 there should be configure check if execinfo.h exists 2014-10-17 10:20:00 otherwise change defined(__linux__) to defined(__GLIBC__) 2014-10-17 10:20:14 for execinfo.h and the place where backtrace() is called 2014-10-17 10:20:19 <[xming]> that's a better option 2014-10-17 10:20:27 +1 2014-10-17 10:34:07 . 2014-10-17 10:34:14 morning 2014-10-17 10:38:31 im trying to move an apkvol from one usb stick to another. how do i get a fresh unconfigured usb of alpine to use the akpvol? 2014-10-17 10:39:32 ScrumpyJack: just put is on it? 2014-10-17 10:40:22 i think it needs untaring 2014-10-17 10:40:28 no 2014-10-17 10:40:30 ScrumpyJack: create a bootable usb from an iso 2014-10-17 10:40:36 *alpine iso 2014-10-17 10:40:39 yeah done that 2014-10-17 10:40:46 it boots fine 2014-10-17 10:40:53 copy the apkovl in the root 2014-10-17 10:40:56 that's it 2014-10-17 10:40:57 done that 2014-10-17 10:41:06 nothing happens 2014-10-17 10:41:36 i have both sticks here. old stick has etc in root 2014-10-17 10:42:26 is the apkovl good? 2014-10-17 10:42:58 boot with usb, then go to /media/usb and untar the apkovl on the / 2014-10-17 10:43:10 ScrumpyJack: it should not ... usb-stick should contain just /apks /boot hostname.apkvol.tar.gz 2014-10-17 10:43:11 yup, it seems fine, I can list contents with lbu when booting the new stick 2014-10-17 10:43:31 cd /media/usb && tar -zxvf $you_apkovl.tar.gz -c / 2014-10-17 10:43:41 hmm, old stick is fat and new stick is ext3 2014-10-17 10:44:08 squashfs treats those very differently i think 2014-10-17 10:44:44 I use ext4 and I just copy apkvol and that´s it ... no need to change anything ... 2014-10-17 10:45:45 that's what I assumed would work 2014-10-17 10:47:54 i'll try again with new a fat fs new stick 2014-10-17 10:49:24 moving the poll.h outside the ifdef and renaming __linux__ to __GLIBC__ makes yadifa compiles. 2014-10-17 10:55:28 i got this yesterday when compiling gammu without glib /usr/include/sys/poll.h:1:2: warning: #warning redirecting incorrect #include to [-Wcpp] 2014-10-17 10:57:51 Is set 2014-10-17 10:58:01 rename it in 2014-10-17 11:00:01 it worked anyway :) 2014-10-17 11:03:37 yes ScrumpyJack 2014-10-17 11:04:08 it works because sys/poll.h has inside include poll.h 2014-10-17 11:18:00 ok, the fs needs to be the same on each stick you move the apkvol to 2014-10-17 11:18:04 bummer 2014-10-17 11:20:26 make sense ... fat has no acl 2014-10-17 11:24:52 <[xming]> fcolista2: okay I have yadifa fully built 2014-10-17 11:27:03 <[xming]> ncopa, barthalion: if you guys have time please review my minetest aport http://sprunge.us/KRUf 2014-10-17 11:29:53 <[xming]> fcolista2: grep -r -l __linux__ * | xargs sed -i -e "s/__linux__/__GLIBC__/g" * plus that #ifndef thing make it compile 2014-10-17 11:30:26 does alpine linux have any plans of using a simpler and leaner init system? 2014-10-17 11:30:37 such as runit, s6, etc 2014-10-17 11:32:04 [xming], yeah, me too. 2014-10-17 11:32:19 <[xming]> I kind of like openrc :D 2014-10-17 11:32:29 +1 2014-10-17 11:32:37 no plans for the moment, diwerth 2014-10-17 11:33:50 that's a pity 2014-10-17 11:44:19 fcolista2, we've been looking at other init systems; but no real plan to switch at this time. 2014-10-17 11:55:41 Perhaps openrc will have supervision some day, which could be considered missing. I'm not fussed about supervision 2014-10-17 12:03:18 <[xming]> supervision would be nice 2014-10-17 12:16:01 diwerth: i have been thinking of runit 2014-10-17 12:16:07 oh he left 2014-10-17 12:18:22 [xming]: minetest aport looks nice 2014-10-17 12:18:28 i'll testbuild it 2014-10-17 12:19:09 Applying: testing/minetest: new aport 2014-10-17 12:19:09 makedepends="sqlite-dev bzip2-dev openssl-dev irrlicht-dev cmake libvorbis-dev 2014-10-17 12:19:09 libjpeg-turbo-dev libpng-dev openal-soft-dev libogg-dev mesa-dev sqlite-dev 2014-10-17 12:19:10 warning: 2 lines add whitespace errors. 2014-10-17 13:43:58 ncopa: hi, about aports, could you please try to pull it ? 2014-10-17 13:49:16 k0r10n: url? 2014-10-17 13:49:43 https://github.com/k0r10n/aports 2014-10-17 13:50:39 looks like you have all the packages in a single commit 2014-10-17 13:50:47 yes, is it bad ? 2014-10-17 13:50:50 https://github.com/k0r10n/aports/commit/aefc6c257918adca022c7f0faddb9a812d1461f9 2014-10-17 13:51:14 you cannot revert parts of it if needed 2014-10-17 13:51:21 but no problem 2014-10-17 13:51:42 it is gerenally better with mutiple minor commits 2014-10-17 13:51:58 i don't have such experience with git, so mayve i do it wrong 2014-10-17 13:52:05 *maybe 2014-10-17 13:52:07 if one of the packages are bad i'll have to reject all 2014-10-17 13:52:45 please notice that i removed all old kde packages 2014-10-17 13:52:53 -arch="" 2014-10-17 13:52:53 +arch="x86 x86_64" 2014-10-17 13:52:59 any reason that arm is disabled? 2014-10-17 13:53:04 akonadi 2014-10-17 13:53:09 yes, i did not test on arm 2014-10-17 13:53:37 i tested only 32 and 64 2014-10-17 13:54:18 when i will learn how to cross compile on alpine, i will test :) 2014-10-17 13:54:29 well, 2014-10-17 13:54:41 currently we normally just set arch="all" 2014-10-17 13:54:46 and thorw it on our arm builder 2014-10-17 13:54:51 and see what happens 2014-10-17 14:03:16 k0r10n: https://github.com/k0r10n/aports/commit/aefc6c257918adca022c7f0faddb9a812d1461f9#diff-6b23acdb622c9c73f0315c744b2804faL5 2014-10-17 14:03:26 when updating pkgver, pkgrel should be reset to 0 2014-10-17 14:04:01 ok 2014-10-17 14:04:22 I would have preferred to fix the test that wrongly detect Backtrace_FOUND: https://github.com/k0r10n/aports/commit/aefc6c257918adca022c7f0faddb9a812d1461f9#diff-6b23acdb622c9c73f0315c744b2804faL5 2014-10-17 14:04:27 since you set #if 0 2014-10-17 14:04:32 that pach cannot be sent upstream 2014-10-17 14:04:55 ah you just ported the patch 2014-10-17 14:05:00 maybe the patch is not needed 2014-10-17 14:05:15 i tested build without it 2014-10-17 14:05:19 but it failed 2014-10-17 14:05:40 ok 2014-10-17 14:06:01 the problem shoudl be reported upstream and the test should be fixed 2014-10-17 14:06:42 https://github.com/k0r10n/aports/commit/aefc6c257918adca022c7f0faddb9a812d1461f9#diff-2f5ea67056953d019d48ed2c371d45a7R7 2014-10-17 14:06:50 ncopa: could I use my existing non-alpine kernel on arm with alpine ... means just put a rootfs on sd and go ? 2014-10-17 14:07:01 would it work with: #if __GLIBC__? 2014-10-17 14:07:17 xen_roger: i would epxect that to work yes 2014-10-17 14:07:38 will give it a try ... ty 2014-10-17 14:08:24 k0r10n: the testing/baloo/no-malloc-xapiandatabase-4.13.3.patch could probably be upstreamed if you used __GLIBC__ instead of __GNUC__ 2014-10-17 14:10:02 with __GLIBC__ i've got only warning that it's not defined, looks like it worked :) 2014-10-17 14:10:20 #if defiend(__GLIBC__) 2014-10-17 14:10:23 is probably more correct 2014-10-17 14:11:07 k 2014-10-17 14:11:22 this testing/kde-workspace/fix-headers-4.11.11.patch looks correct and could probably be upstreamed 2014-10-17 14:12:36 this too: testing/kde-workspace/uint-fadeeffect-4.11.11.patch 2014-10-17 14:12:58 +command="/usr/bin/kdm -nodaemon" 2014-10-17 14:13:03 does that work= 2014-10-17 14:13:04 ? 2014-10-17 14:13:13 yes :) 2014-10-17 14:13:18 i rhink you can do: command_args="-nodaemon" 2014-10-17 14:13:47 but it's better to pass arguments in other var.... 2014-10-17 14:14:22 i tested on virtual machine and it started without error 2014-10-17 14:14:45 ncopa: any doc on how to create rootfs of alpine for arm ? debian I used debootstrap and qemu static for stage2 2014-10-17 14:14:54 this looks hackish: testing/kdelibs/fix-cmake-FindLibintl-4.13.3.patch 2014-10-17 14:15:07 it looks like the code after is supposed to detect libintl 2014-10-17 14:15:29 xen_roger: apk add --root /path/to/root --aech 2014-10-17 14:15:33 ncopa: yes, because without it it did not want to find dgettext.. 2014-10-17 14:15:35 xen_roger: apk add --root /path/to/root --arch armhf 2014-10-17 14:15:36 i think 2014-10-17 14:16:17 ncopa: and linking with libintl failed.... 2014-10-17 14:17:10 yes, i dont doubt that, but i think the patch is not the "correct" (tm) way to fix it 2014-10-17 14:17:16 so it cannot be upstreamed 2014-10-17 14:17:22 :/ 2014-10-17 14:17:30 works for us though 2014-10-17 14:17:36 k 2014-10-17 14:17:50 i htink we want make patches in such way that we can upstream them 2014-10-17 14:17:54 as far as possible 2014-10-17 14:18:11 i think we can upstream this too: testing/kdelibs/fix-headers-kdelibs.patch 2014-10-17 14:18:15 it looks correct 2014-10-17 14:18:36 no 2014-10-17 14:18:45 +-#include // ino_t 2014-10-17 14:18:45 ++#include // ino_t , but not there !! 2014-10-17 14:18:45 ++#include // maybe, here ? 2014-10-17 14:19:07 we can probably just replace sys/param.h with sys/types.h 2014-10-17 14:19:30 if its ino_t that we need 2014-10-17 14:19:31 then 2014-10-17 14:19:37 http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/sys_stat.h.html 2014-10-17 14:19:46 sys/stat.h is what we should use 2014-10-17 14:20:09 +pkgrel=1 2014-10-17 14:20:14 testing/kdepimlibs/APKBUILD 2014-10-17 14:20:20 we start counting from 0 2014-10-17 14:20:56 testing/libxss/APKBUILD 2014-10-17 14:21:03 +depends="scrnsaverproto" 2014-10-17 14:21:03 +depends_dev="util-macros libxext-dev" 2014-10-17 14:21:18 i would expect that scrnsaverproto should go to makedepends only 2014-10-17 14:21:25 we probably don need it runtime 2014-10-17 14:21:35 yes sure 2014-10-17 14:22:30 +#-DCMAKE_SKIP_RPATH=ON "$_builddir" || return 1 2014-10-17 14:22:35 we coudl just remove that line 2014-10-17 14:22:41 keep things clean and tidy 2014-10-17 14:22:55 ...., sry forgot to clean it.. 2014-10-17 14:23:06 np, i work that way too :) 2014-10-17 14:25:26 soprano has not dbus support? 2014-10-17 14:25:35 testing/soprano/soprano-2.9.4-dbus-1.patch 2014-10-17 14:25:52 soprano is just for testing, because i built all without soprano and nepomuk 2014-10-17 14:26:06 ok 2014-10-17 14:26:14 but really 2014-10-17 14:26:19 this is good 2014-10-17 14:26:26 overall good 2014-10-17 14:26:32 good job! 2014-10-17 14:26:33 ok, will try it 2014-10-17 14:26:49 i would have preferred to have each package in separate git commit 2014-10-17 14:26:51 but its ok 2014-10-17 14:27:02 for this time ;) 2014-10-17 14:27:19 i can do it, but it will take many, many, many time 2014-10-17 14:27:26 next time 2014-10-17 14:28:11 you did an impressive amount of work there 2014-10-17 14:28:23 and it seems that it builds too 2014-10-17 14:31:10 ty, but about libintl in kdelibs, it failed because it tried to link without -lintl, and this patch only adds this flag 2014-10-17 14:38:45 k0r10n: soprano fails to build due to missing iodbc 2014-10-17 14:38:52 and libiodbc fails to build on my box 2014-10-17 14:39:42 ok 2014-10-17 14:39:51 soprano is not really needed i undersand? 2014-10-17 14:39:52 ok 2014-10-17 14:39:57 i push it as is 2014-10-17 14:40:07 thanks 2014-10-17 14:40:07 yes, but i can submit a patch for libiodbc if u need 2014-10-17 14:40:19 Linux User 2014-10-17 14:40:25 lol 2014-10-17 14:40:32 ty 2014-10-17 14:40:48 Re: supervision, I had some success experimenting with s6 the other day. I can throw my packages up on github if anyone's interested in taking a look. 2014-10-17 14:43:08 jlyo: i kinda like the concept with s6 but i think the general implmentation is a bit 'weird' 2014-10-17 14:43:20 the djb coding style 2014-10-17 14:45:02 but i kinda feel for writing our own supervision manager, since i dont think any of them gives what we want 2014-10-17 14:45:06 dont have time unfortunally 2014-10-17 14:45:23 Yeah, he does his own thing. I like the exec() chainloading though. 2014-10-17 14:46:12 exec chainloading is kinda interesting 2014-10-17 14:46:25 i dont htink i like the fit the entire script in argv 2014-10-17 14:47:14 but i like much of the reasoning 2014-10-17 14:49:20 ncopa: libiodbc - http://sprunge.us/gYFa 2014-10-17 14:50:59 It'd be easy to pick apart components. I think a quick way to get it started would be to unshare the PID namespace so svscan gets PID 1 local to it's children. 2014-10-17 14:54:18 k0r10n: thanks 2014-10-17 14:59:58 ncopa: what does options make ? 2014-10-17 15:00:17 i did find it in doc 2014-10-17 15:01:41 its in my local abuild 2014-10-17 15:01:46 new abuild feature 2014-10-17 15:02:04 basically, it will exit with error if suidbit set on executable 2014-10-17 15:02:16 ok i need to go 2014-10-17 15:02:25 have a nice weekend 2014-10-17 15:02:32 ty, u too 2014-10-17 17:56:31 ncopa: we are discussing the points you mentioned regarding the syslog-ng internals in the current testing package 2014-10-17 17:57:41 especially the parts regarding the runlevel modification during removal of the package 2014-10-17 17:58:02 apk has a lot of functionality for such cases but they should not be used 2014-10-17 17:58:05 why??? 2014-10-17 17:58:30 if a package is going to be removed, then it should remove itself from the configured runlevels 2014-10-17 17:58:39 or in other words: 2014-10-17 17:59:02 if there are such possibilities like pre- and post-install and pre- and post-uninstall scripts, 2014-10-17 17:59:07 why not use them for such things? 2014-10-17 17:59:27 this is a common use case and we can't see the point in not using them 2014-10-17 18:01:05 my colleague just send a patch some minutes ago with a lot of changes but we think there are some changes that should be handled in a different/better way 2014-10-17 18:09:03 actually removing a service from runlevel sounds sane 2014-10-17 18:41:15 barthalion: why? 2014-10-17 18:41:30 if a package get removed, it should be removed from the used runlevels 2014-10-17 18:41:42 otherwise you got errors on next reboot 2014-10-17 18:44:35 ah, probably misunderstood you 2014-10-17 18:44:49 so you think it is ok to remove it during uninstall or not? 2014-10-17 19:25:34 StarWarsFan: I think it's ok 2014-10-18 09:34:18 <[xming]> when a package only provides .h and .a how should I package it? Will normal way it will produce a pkg with only the dir /usr and a -dev pkg 2014-10-18 09:34:34 <[xming]> should I name the pkg something-dev? 2014-10-18 10:25:49 well, I guess it's ok to just keep $pkgname 2014-10-18 10:26:18 what library we're speaking about? 2014-10-18 11:12:09 <[xming]> it's libevhtp 2014-10-18 11:13:00 did someone push your minetest patch yet? 2014-10-18 11:14:14 I'll take a look at libevhtp, I need to make a dinner first 2014-10-18 11:14:52 but at the first glance, I'd just keep libevhtp name, no suffixes 2014-10-18 11:15:20 if you don't put $pkgname-dev in subpackages, files should stay within the main pkg 2014-10-18 11:18:04 <[xming]> barthalion: nacioa pushed it 2014-10-18 11:18:43 <[xming]> so there is no convention that headers + static libs should be in a -dev pkg? 2014-10-18 11:20:08 <[xming]> it was ncopa :D 2014-10-18 11:20:43 Is anything wrong with my i3lock aport? http://lists.alpinelinux.org/alpine-devel/4028.html 2014-10-18 11:20:57 <[xming]> am I trying to have a seafile aport, it's a nightmare with dependencies 2014-10-18 11:22:39 <[xming]> jomat: +source="$url/$pkgname-$pkgver.tar.bz2" # cflags.patch" -> huh? 2014-10-18 11:23:19 <[xming]> it looks alright to me, but hey who am I 2014-10-18 11:23:39 I'll recheck that again, thanks :-) 2014-10-18 11:25:22 No, no cflags.patch needed here :-) This comment can be removed :-) 2014-10-18 11:26:57 <[xming]> jomat: also ncopa told me that we need proper commit msg 2014-10-18 11:27:53 <[xming]> testing/i3lock: new aport 2014-10-18 11:28:11 <[xming]> An improved screenlocker based upon XCB and PAM 2014-10-18 11:28:19 <[xming]> http://i3wm.org/i3lock/ 2014-10-18 11:28:31 <[xming]> like this http://git.alpinelinux.org/cgit/aports/commit/?id=276617c77708f50d68ba9b4c87a866399ddd7250 2014-10-18 11:30:57 <[xming]> anyone working on one of these gunicorn-python python-djblets python-flup python-webpy python-chardet 2014-10-18 11:39:46 What's the right way to track/update packages to update their library dependencies? In my case, the quicktun packages on the mirrors needs libsodium.so.10, but the current version is libsodium.so.13 ... if i just rebuild the package from the aports tree, the dependency is updated on my local system and can be installed. 2014-10-18 11:40:26 someone bumped libsodium and forgot to rebuild dependent packages 2014-10-18 11:40:50 ls 2014-10-18 11:40:52 oops 2014-10-18 11:41:20 jomat: are we talking about edge? 2014-10-18 11:42:47 barthalion: testing 2014-10-18 11:44:11 libsodium@testing has usr/lib/libsodium.so.13, quicktun@testing requires so:libsodium.so.10 (missing) 2014-10-18 11:44:28 I'll fix it in a moment, let me test the build first 2014-10-18 11:44:37 thx :-) 2014-10-18 11:44:51 my build environment is quite out of date 2014-10-18 11:55:24 jomat: ↑ 2014-10-18 11:55:51 apk update && apk add quicktun@testing should do it, shouldn't it? 2014-10-18 11:56:19 so should apk add -U quicktun@testing ;) 2014-10-18 11:56:29 [xming]: well, there is convention, but as long as package provides only headers and static files, why bother? 2014-10-18 11:57:22 Ah, how long does it take until the mirrors are synced? I use dl-3.alpinelinux.org :-) 2014-10-18 11:58:24 an hours or so 2014-10-18 11:58:34 nl.alpinelinux.org gets packages almost instantly 2014-10-18 11:58:45 s/hours/hour/ 2014-10-18 11:59:24 <[xming]> barthalion: just to be cosmetically corrent? 2014-10-18 11:59:39 <[xming]> correct* 2014-10-18 12:00:24 when I see a foo-dev package, I expect to see the foo as well 2014-10-18 12:00:37 this is not the case here 2014-10-18 12:01:17 your foo-dev won't be a subpackage of foo, it would be splitted from nowhere 2014-10-18 12:01:35 but as always if in doubt, wait for fabled or ncopa and let them decide 2014-10-18 12:44:06 <[xming]> okay that makes sense too 2014-10-18 13:06:20 hi 2014-10-18 13:06:53 I could be able to make use of automatic mounting in XFCE 2014-10-18 13:07:06 udev+udisks+udisks2+gvfs is needed 2014-10-18 13:07:12 but now I have issues with ntfs-3g 2014-10-18 13:07:29 when I try to mount NTFS partitions 2014-10-18 13:07:35 it can't 2014-10-18 13:08:04 it wants to run this command 2014-10-18 13:08:37 mount -t "ntfs" -o "nosuid,uid=1000,gid=1000,dmask=0077,fmask=0177" "/dev/sda1" "/run/media/..." 2014-10-18 13:08:40 but I get 2014-10-18 13:08:55 NTFS-fs error (device sda1): parse_options(): Unrecognized mount option uhelper. 2014-10-18 13:08:58 in kernel messages 2014-10-18 13:09:24 by removing uhelper from the above command 2014-10-18 13:09:30 sorry! I removed it already 2014-10-18 13:09:33 it works fine 2014-10-18 13:09:37 So I want to knwo 2014-10-18 13:09:38 know 2014-10-18 13:09:48 where can I change that command 2014-10-18 13:09:51 is it in UDEV 2014-10-18 13:10:01 or is it in UDISKS2 2014-10-18 13:10:04 or where? 2014-10-18 13:10:42 this is the original command it wants to run 2014-10-18 13:10:58 mount -t "ntfs" -o "uhelper=nosuid,uid=1000,gid=1000,dmask=0077,fmask=0177" "/dev/sda1" "/run/media/..." 2014-10-18 13:11:04 any idea? 2014-10-18 13:11:47 <[xming]> there is no udev nor spoon 2014-10-18 13:12:33 so is it the problem with mount? 2014-10-18 13:12:41 I mean busybox mount 2014-10-18 13:13:08 <[xming]> there is no udev in Alpine 2014-10-18 13:13:09 I have installed udev and udisks to achieve automounting 2014-10-18 13:13:16 <[xming]> ah 2014-10-18 13:13:21 apk add udev 2014-10-18 13:13:33 This is not a problem it seems 2014-10-18 13:13:36 <[xming]> then probably it's bb's mount 2014-10-18 13:13:51 I've installed udev+udisks+udisks2+gvfs+ntfs3g 2014-10-18 13:13:53 all work fine 2014-10-18 13:14:17 yeah I guess so 2014-10-18 13:14:29 maybe I can use a workaroun 2014-10-18 13:14:32 workaround 2014-10-18 13:14:36 create a script 2014-10-18 13:14:41 that drops this option 2014-10-18 13:14:50 how's that? 2014-10-18 13:15:01 <[xming]> chatr: have you installed util-linux? 2014-10-18 13:15:07 no 2014-10-18 13:15:37 yeah man! 2014-10-18 13:15:38 that fixed it 2014-10-18 13:15:40 :-)) 2014-10-18 13:15:43 great job! 2014-10-18 13:15:46 <[xming]> try that first, see if the mount in that accpets uhelper 2014-10-18 13:15:56 <[xming]> okay I am slow :( 2014-10-18 13:16:06 I'll write my experience for the Alpine wiki 2014-10-18 13:16:19 for helping set up XFCE 2014-10-18 13:16:31 ur cool man :-) 2014-10-18 13:17:07 <[xming]> if you write a wiki then you are cool too 2014-10-18 13:17:31 :-) 2014-10-18 13:17:45 So, let me ask more! 2014-10-18 13:17:54 is there a way to see apk files? 2014-10-18 13:18:05 like dpkg -L 2014-10-18 13:18:07 or so 2014-10-18 13:18:25 is it zip or something? 2014-10-18 13:18:42 I want to keep track of installed files 2014-10-18 13:18:57 for example, I want to know what util-linux has installed 2014-10-18 13:18:59 apk info -l 2014-10-18 13:19:02 <[xming]> it's a tgz 2014-10-18 13:19:02 ehhr 2014-10-18 13:19:05 apk info -L 2014-10-18 13:19:12 great 2014-10-18 13:19:13 :-) 2014-10-18 13:19:22 I don't have man 2014-10-18 13:19:28 so you couldn't tell RTFM 2014-10-18 13:19:31 :-D 2014-10-18 13:19:38 <[xming]> man man 2014-10-18 13:19:48 apk add man? 2014-10-18 13:19:49 browse content of an .apk: tar -tzf foo.apk 2014-10-18 13:20:11 I don't know where it is downloaded 2014-10-18 13:20:31 apk add man 2014-10-18 13:20:48 <[xming]> as jomat said, apk info -L works 2014-10-18 13:20:49 didn't install man, even for itself 2014-10-18 13:20:57 yeah apk info -L worked just fine 2014-10-18 13:21:15 <[xming]> apk add man-pages 2014-10-18 13:21:33 I've just install man binary only! 2014-10-18 13:21:33 <[xming]> you should 've man man first to find all this out :p 2014-10-18 13:21:34 :-) 2014-10-18 13:21:40 <[xming]> ACTION ducks 2014-10-18 13:21:53 man man > No manual entry for man 2014-10-18 13:22:05 one should bundle man man with itself 2014-10-18 13:22:06 :-/ 2014-10-18 13:22:09 <[xming]> I was just teasing you :( 2014-10-18 13:22:10 apk add man-doc 2014-10-18 13:22:12 let's bootstrap! 2014-10-18 13:22:23 <[xming]> not doc, you need man-pages 2014-10-18 13:22:55 yeah 2014-10-18 13:23:03 But the manpage for man is in man-doc 2014-10-18 13:23:05 man man works with man-pages even when I removed man-doc 2014-10-18 13:23:12 Oh 2014-10-18 13:23:38 apk info -L man-doc 2014-10-18 13:23:40 empty! 2014-10-18 13:23:56 intriguing 2014-10-18 13:23:56 <[xming]> :D 2014-10-18 13:24:13 well, it's obviously empty if you removed it 2014-10-18 13:24:19 <[xming]> it's pulling tha man page from /dev/urandom when you use man man 2014-10-18 13:25:34 barthalion, you're right 2014-10-18 13:26:06 how odd! 2014-10-18 13:26:15 this packages have conflict 2014-10-18 13:26:22 there's no problem? 2014-10-18 13:26:47 wait 2014-10-18 13:27:07 this one is from man-pages 2014-10-18 13:27:11 usr/share/man/man1/man.1p 2014-10-18 13:27:15 this one is from man-doc 2014-10-18 13:27:18 usr/share/man/man1/man.1 2014-10-18 13:27:27 so maybe there's no conflict 2014-10-18 13:28:28 1p is POSIX manual 2014-10-18 13:28:57 how can I see that? 2014-10-18 13:29:35 <[xming]> man 1p man 2014-10-18 13:30:20 this worked for me: 2014-10-18 13:30:26 man man 1p 2014-10-18 13:30:31 oh nwo 2014-10-18 13:30:33 no 2014-10-18 13:30:37 that's man man 2014-10-18 13:30:38 :-( 2014-10-18 13:30:57 $man 1p man> No entry for man in section 1p of the manual 2014-10-18 13:32:07 <[xming]> strange man 2014-10-18 13:33:46 this shows that in section 7 2014-10-18 13:33:49 man man -a 2014-10-18 13:34:02 and then pressing q to exit first one: man(1) 2014-10-18 13:34:36 man 7 man 2014-10-18 13:34:38 :-) 2014-10-18 13:34:39 it worked 2014-10-18 13:34:58 but am I correct? 2014-10-18 13:35:00 1p=7? 2014-10-18 13:35:58 it seems no 2014-10-18 13:36:00 :-/ 2014-10-18 13:36:46 tihs works: man /usr/share/man/man1/man.1p 2014-10-18 13:36:57 *this 2014-10-18 13:40:50 btw 2014-10-18 13:41:01 how do you compare performance of Alpine 2014-10-18 13:41:07 to other distros? 2014-10-18 13:41:15 how do you measure boot time? 2014-10-18 13:41:39 how do you measure application startup time? 2014-10-18 13:41:56 in runtime 2014-10-18 13:42:10 without profiling tools? 2014-10-18 13:49:49 tnx barthalion 2014-10-18 13:50:00 tnx [xming] 2014-10-18 13:50:06 bye 2014-10-18 13:50:21 tnx jomat 2014-10-18 13:51:06 cu 2014-10-18 14:18:21 wtf? 2014-10-18 15:41:14 hi again 2014-10-18 15:41:19 I want to use ALSA 2014-10-18 15:41:34 but I don't know which kernel module I should load 2014-10-18 15:41:41 I looked at what UBUNTU does 2014-10-18 15:42:04 there's a long list of possible modules in modprobe.d 2014-10-18 15:42:15 for alsa-base.conf in UBUNTU 2014-10-18 15:42:41 I can enable OSS as described in Alpine FAQ 2014-10-18 15:42:46 but I can't enable ALSA 2014-10-18 15:42:51 and most of apps use ALSA 2014-10-18 15:43:35 I even dropped UBUNTU version of modprobe.d/alsa-base.conf 2014-10-18 15:43:45 in /etc/modprobe.d of Alpine Linux 2014-10-18 15:43:52 but it didn't help 2014-10-18 15:45:09 hm 2014-10-18 15:45:17 now that I use lsmod|grep snd 2014-10-18 15:45:26 I see several modules inserted 2014-10-18 15:45:40 but even now alsamixer says 2014-10-18 15:45:46 cannot open mixer: No such file or directory 2014-10-18 15:46:13 wow 2014-10-18 15:46:17 it works fine in root 2014-10-18 15:46:25 why it doesn't work for normal user? 2014-10-18 15:47:49 maybe it's a permission problem 2014-10-18 15:47:55 but should I change permissons? 2014-10-18 15:55:22 hm 2014-10-18 15:55:32 I should add myself to audio group 2014-10-18 15:55:38 but as far as I remember 2014-10-18 15:56:00 I couldn't add a user to some group here 2014-10-18 15:56:01 yeah, basically i had to add myself to the audio group 2014-10-18 15:56:14 adduser chatr audio 2014-10-18 15:56:43 tnx 2014-10-18 15:58:33 groups says I am not in audio groups 2014-10-18 15:58:55 but when I use "groups chatr" in root 2014-10-18 15:59:04 it says I am in the audio group 2014-10-18 15:59:12 maybe I should log out/in? 2014-10-18 15:59:47 yeah 2014-10-18 15:59:50 su chatr worked 2014-10-18 15:59:52 u have to logout and then back in 2014-10-18 15:59:55 I'm trying it 2014-10-18 16:00:01 will be back soon 2014-10-18 16:31:03 ok 2014-10-18 16:31:05 it worked fine 2014-10-18 16:31:09 and now I have audio 2014-10-18 16:33:23 but as I found 2014-10-18 16:33:34 here there's no way to connect to WIFI in GUI 2014-10-18 16:33:35 http://wiki.alpinelinux.org/wiki/Connecting_to_a_wireless_access_point 2014-10-18 16:33:38 am I right? 2014-10-18 16:34:48 Xubuntu has a GUI for this 2014-10-18 16:35:13 I'm trying to make things more simple 2014-10-18 16:35:32 I know iwconfig and ifconfig are great tools 2014-10-18 16:35:33 :-) 2014-10-18 16:35:57 and also wpa_supplicant 2014-10-18 16:37:27 is there a chance 2014-10-18 16:37:40 that I can search in all packages? 2014-10-18 16:37:46 Ubuntu has packages.ubuntu.com 2014-10-18 16:37:57 that lets you to find files in all packages 2014-10-18 16:38:03 for example I want lshw 2014-10-18 16:38:08 where should I search? 2014-10-18 16:38:39 It seems that it's available in testing 2014-10-18 16:38:41 :-) 2014-10-18 16:38:44 but anyway 2014-10-18 16:39:20 it's great to have such thing in http://alpinelinux.org/packages 2014-10-18 16:46:17 chatr: i have wicd in TODO list, maybe will be available soon.. 2014-10-18 16:46:34 tnx 2014-10-18 16:46:50 k0r10n, don't you think an XFCE based image could be a good idea? 2014-10-18 16:46:59 I'm trying to reach a good XFCE configuration 2014-10-18 16:47:14 you can try KDE :) 2014-10-18 16:47:15 with the goal of releasing such image 2014-10-18 16:47:17 :-)) 2014-10-18 16:47:37 or even Windows 8 2014-10-18 16:47:39 :-D 2014-10-18 16:47:43 10 maybe 2014-10-18 16:47:58 APKBUILDS for KDE are ready, there are waiting to build :) 2014-10-18 16:48:12 *they 2014-10-18 16:48:16 Thought you were kidding! 2014-10-18 16:49:02 LXQt can also be a good choice 2014-10-18 16:49:10 i also would like to make KDE based Alpine image 2014-10-18 16:49:18 that would be great! 2014-10-18 16:49:27 if i will have time, i will do 2014-10-18 16:50:10 i tested it on Virtualbox with 512 Mb ram - works very fast :)) 2014-10-18 16:50:20 wow 2014-10-18 16:50:26 what about installation? 2014-10-18 16:50:33 right now, one needs to load an image, then install it for example on a USB stick 2014-10-18 16:50:54 why not create an image, that is easily deployed on a USB stick? 2014-10-18 16:51:18 yes, image is good idea 2014-10-18 16:51:29 I'm working on a USB stick 2014-10-18 16:51:36 then I can make a dd image of it 2014-10-18 16:51:40 and then gzip/bzip it 2014-10-18 16:51:52 would it be good for deployment? 2014-10-18 16:51:57 will your image have XFCE ? 2014-10-18 16:52:03 yes 2014-10-18 16:52:11 that's I'm working on 2014-10-18 16:52:16 you mean for installing ? 2014-10-18 16:52:20 no 2014-10-18 16:52:26 an image 2014-10-18 16:52:32 made from the USB stick I'm working on 2014-10-18 16:52:50 I don't call it installation 2014-10-18 16:52:58 a dd 2014-10-18 16:53:02 like live USB ? 2014-10-18 16:53:03 and you're there! 2014-10-18 16:53:05 yes 2014-10-18 16:53:36 do you use alpine-iso to make image ? 2014-10-18 16:53:52 I'm not familiar with that 2014-10-18 16:54:17 it's simple - just makefiles 2014-10-18 16:55:30 http://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image 2014-10-18 16:55:46 tnx 2014-10-18 16:56:46 I have exhausted ~1Gig right now 2014-10-18 16:56:47 :-) 2014-10-18 16:56:54 So, that will be a big image 2014-10-18 16:57:23 ) 2014-10-18 16:58:15 working with this custom ISOs is easier than dd'ed image 2014-10-18 16:58:17 isn't it? 2014-10-18 16:59:04 yes, much easier 2014-10-18 16:59:29 just make PROFILE=your-image iso 2014-10-18 17:00:25 What about default settings of apps? 2014-10-18 17:00:40 for example, there's a bug in default Gtk themes 2014-10-18 17:00:52 So, I have to install a custom Gtk3 theme 2014-10-18 17:01:08 how's possible to mange such things 2014-10-18 17:02:05 I should rebuild apks maybe? 2014-10-18 17:02:14 you can use apkovl 2014-10-18 17:02:54 local caches 2014-10-18 17:03:00 with updated things 2014-10-18 17:03:02 isn't it? 2014-10-18 17:03:32 something like that 2014-10-18 17:03:50 you can see example in alpine-desktop conf.mk file 2014-10-18 17:04:03 and download alpine-live.apkovl.tar.gz 2014-10-18 17:04:08 and look inside 2014-10-18 17:04:31 in fact apkovl is just archive created with tar 2014-10-18 17:08:26 hm 2014-10-18 17:08:42 I didn't knew wpa_supplicant has a gui 2014-10-18 17:09:00 this is useful for WIFI 2014-10-18 17:09:09 and users will want it for sure 2014-10-18 17:09:31 I'm including Qt for VLC 2014-10-18 17:09:47 so qt dependencies would be no problem 2014-10-18 17:10:43 Maybe I should add it to XFCE startup 2014-10-18 17:10:54 because it's a standalone application 2014-10-18 17:10:59 and not a panel plugin 2014-10-18 17:11:50 i will wait for your image :), gotta go, c u 2014-10-18 17:11:56 :-) 2014-10-18 17:11:58 tnx 2014-10-18 17:11:59 bye 2014-10-18 17:58:37 k0r10n: what's the rationale behind lowering akonadi's pkgrel? 2014-10-19 15:40:46 barthalion: about akonadi, because ncopa sad if new pkgver, pkgrel should be set to 0 2014-10-19 16:00:49 pkgver doesn't look to be changed in your patch 2014-10-19 16:10:41 it's because last commit rewrite it from 1.12.1 to 1.13 2014-10-19 16:13:53 to be exact from 1.5.3 to 1.13.0 2014-10-19 16:14:14 http://git.alpinelinux.org/cgit/aports/commit/?id=9085b73f09d41af164193a3e0b5e22fcce80d503 2014-10-19 16:15:57 well, it's too late to change pkgrel now IMHO 2014-10-19 16:16:04 especially make it lower 2014-10-19 16:16:35 as clear upgrade won't be possible, as 0 is "older" than 3 2014-10-19 16:16:50 :/ 2014-10-19 18:06:36 not that we care much about testing :D 2014-10-19 19:05:06 hi guys 2014-10-19 19:05:50 Today I was thinking about musl libc performance 2014-10-19 19:05:59 why is that superior to glibc? 2014-10-19 19:06:43 I understand uClibc take less space 2014-10-19 19:07:19 dietlibc I meant 2014-10-19 19:07:37 because it "only implements the most important and commonly used functions." 2014-10-19 19:07:54 but why musl is smaller in binary 2014-10-19 19:07:58 and faster in runtime? 2014-10-19 19:08:15 why is that only available for Linux? 2014-10-19 19:08:54 is it faster because Linux system calls? 2014-10-19 19:09:13 well 2014-10-19 19:09:19 Every library on Linux should directly/indirecty use Linux system calls 2014-10-19 19:09:21 isn't it? 2014-10-19 19:09:22 we do not use glibc because we just don't want to 2014-10-19 19:09:35 any other reasons for using it would have to go to #musl 2014-10-19 19:09:39 I want to understand the difference 2014-10-19 19:09:45 why is musl faster? 2014-10-19 19:09:48 is it really? 2014-10-19 19:10:05 I didn't measure and compare performance. 2014-10-19 19:11:24 But Alpine with XFCE worked blazingly fast 2014-10-19 19:11:26 on my system 2014-10-19 19:11:40 which I tested before Xubuntu on it 2014-10-19 19:11:52 It was good 2014-10-19 19:12:06 But Alpine provided visible performance improvement 2014-10-19 19:12:14 at least it was what I was feeling 2014-10-19 19:12:33 XFCE windows opened nearly instantly 2014-10-19 19:12:38 when I clicked on them 2014-10-19 19:12:52 chatr: only a small part of libc is about the raw syscalls 2014-10-19 19:13:22 nsz, So, does it really improves performance? (I feel so) 2014-10-19 19:13:29 Why it's faster? 2014-10-19 19:13:48 Is there a particular reason for that? 2014-10-19 19:14:14 musl is much more careful about not adding useless dependencies 2014-10-19 19:15:41 the glibc dynamic loader is slower than the musl one and glibc will try to look for locale files at various places and does some useless administration even when it is not needed etc 2014-10-19 19:15:47 there are lot of details 2014-10-19 19:17:52 Alpine gives me a feel of real time 2014-10-19 19:18:19 I don't know if musl guys 2014-10-19 19:18:26 the number of syscalls at program startup are much lower with musl just because it does not try to open 100 useless config files 2014-10-19 19:18:40 have used design and implementation principles of realtime systems 2014-10-19 19:20:12 So, considering GLIBC compatibility, theoretically there shouldn't be obstacles for recompiling all packages of for exmple Ubuntu with musl 2014-10-19 19:20:34 Is that true? 2014-10-19 19:20:46 yes, except binaryware 2014-10-19 19:20:56 like opengl drivers, flash plugins etc 2014-10-19 19:21:11 even those could work because of binary compat, but they need some work 2014-10-19 19:21:18 yeah 2014-10-19 19:22:27 there are some other obstacles 2014-10-19 19:22:40 like legacy setups with weird encoing musl wont support that 2014-10-19 19:23:30 or the nss plugin system for namelookups wont work with musl 2014-10-19 19:23:47 and various other bits 2014-10-19 19:24:36 those usually have saner alternatives than depending on libc internals, but historically many systems depend on glibc behaviour 2014-10-19 19:26:54 actually we are not sure why gui programs would be faster on musl but ppl reported that they feel faster 2014-10-19 19:27:59 I feel so 2014-10-19 19:28:07 probably 'first startup time' is the reason (reading stuff from disk can be slow and glibc tends to do more of it) 2014-10-19 19:28:09 It doesn't have that annoying lags 2014-10-19 19:28:29 I hate that! 2014-10-19 19:28:37 System works fine 2014-10-19 19:28:47 But randomly waits for a while 2014-10-19 19:29:03 that jitter bothers user 2014-10-19 19:29:20 I don't see this kind of jitter here 2014-10-19 19:29:40 Still some apps like Firefox have that 2014-10-19 19:30:05 But it's not a common symptom! 2014-10-19 19:30:33 As a user, I feel that Chrome is superior to Firefox in this sense 2014-10-19 19:30:58 I guess, several layers of software causes this 2014-10-19 19:31:08 When you're close to metal 2014-10-19 19:31:20 You'll have less issues in this sense 2014-10-19 19:31:38 Let's add garbage collection stuff to this 2014-10-19 19:32:03 For example in Java 2014-10-19 19:32:11 we have this issue 2014-10-19 19:35:33 A question about systemd 2014-10-19 19:35:57 Although Alpine provides superior speed in apps (at least this is wha I feel) 2014-10-19 19:36:07 its boot time is not so much better 2014-10-19 19:36:21 than for example 2014-10-19 19:36:23 Ubuntu 2014-10-19 19:36:35 Why is that? 2014-10-19 19:36:41 I'm not a fan of Systemd 2014-10-19 19:37:04 but will/can systemd speed up boot sequence? 2014-10-19 19:38:10 How's OpenRC compared to systemd? 2014-10-19 19:51:01 Interesting 2014-10-19 19:51:22 https://www.openhub.net/p/musl > representing 83,915 lines of code 2014-10-19 19:51:36 https://www.openhub.net/p/glibc > representing 2,379,265 lines of code 2014-10-19 19:52:06 jaw-dropping! 2014-10-19 19:54:35 I can't explain that, unless I accept that there's only more important things available on musl. 2014-10-19 19:55:09 "dietlibc was developed from scratch and thus only implements the most important and commonly used functions." 2014-10-19 19:55:39 but: https://www.openhub.net/p/8181 > representing 76,901 lines of code 2014-10-19 19:56:19 and it seems both musl and dietlibc arise from earlier versions of GLIBC. 2014-10-19 19:56:22 Is that correct? 2014-10-19 19:56:28 no 2014-10-19 19:57:00 musl is not? 2014-10-19 19:57:05 they have no connection to glibc in any way (otherwise they would have to use the same license) 2014-10-19 20:00:57 tnx man 2014-10-19 20:01:01 I should go now 2014-10-19 20:01:32 nsz, bye. 2014-10-19 21:54:28 Hello eveyone, i'm trying to build "collectd-libvirt" because it's not in the collectd package(s). For this i installed a vserver (alpine 2.7) and 2 vserver guests (2.7 and 3.0). So far i have no luck, the configure from collectd won't activate libvirt apparently because of an issue with libintl. I installed the libvirt-dev package. Can anyone help ? my goal is to have it work in 2.7 as i have an Alpine 2.7.9 running as a Xen host that i prefer not to 2014-10-19 21:54:28 upgrade because i dont want to break my current setup. 2014-10-19 22:06:42 hi pierre76 2014-10-19 22:10:05 hi 2014-10-19 22:10:15 it seems i cannot build collectd on edge at all 2014-10-19 22:10:41 i see… i was trying on 3.0.4 2014-10-19 22:11:00 and you could compile it? 2014-10-19 22:11:31 i think i could before trying to add the libvirt plugin, let me retry 2014-10-19 22:12:21 ok after rebuild libvirt i can 2014-10-19 22:15:45 pierre76: collectd in edge has libvirt-dev included 2014-10-19 22:16:06 so i think it already has libvirt support. im not sure about 2.7 though 2014-10-19 22:17:26 yes, this it has but it does not create a collectd-libvirt subpackage ? 2014-10-19 22:19:03 no, its not defined 2014-10-19 22:19:13 so its probably inside collectd by default 2014-10-19 22:20:00 im not sure about 3.0, but it looks like libvirt needs a rebuild on edge. i think its because of recent xen upgrade or something like that. 2014-10-19 22:20:56 checking libvirt/libvirt.h usability... yes 2014-10-19 22:20:56 checking libvirt/libvirt.h presence... yes 2014-10-19 22:20:56 checking for libvirt/libvirt.h... yes 2014-10-19 22:20:56 checking for virDomainBlockStats in -lvirt... yes 2014-10-19 22:21:28 its listed as a plugin in collectd which is why i expected a subpackage 2014-10-19 22:21:40 https://collectd.org/wiki/index.php/Plugin:libvirt 2014-10-19 22:22:14 im not sure what the reason is why its not a subpkg 2014-10-19 22:22:23 i didnt build/manage that pkg. 2014-10-19 22:22:48 if you like it to have a subpkg, you can send us a patch or else report it on bugs.a.o 2014-10-19 22:25:24 i'll try by myself to port it back to 2.7 see if it works as is and if i cant i'll see about a bug/feature 2014-10-19 22:25:39 ncopa is the maintainer :) 2014-10-19 22:27:03 pierre76: it should be build with libvirt also in 2.7 2014-10-19 22:27:15 alteast thats what git tells me. 2014-10-19 22:28:27 can you tell me how to rebase to 2.7 ? 2014-10-19 22:35:11 pierre76: what do you mean? 2014-10-19 22:37:49 i mean compile this for 2.7 2014-10-19 22:39:32 i'm not used to managing branches and compile for an older version 2014-10-19 23:02:44 ok managed it, thanks for the help clandmeter, i'll try a few more things to see it collectct-libvirt works as is, or if it needs a package or anything 2014-10-20 06:37:05 morning 2014-10-20 06:37:19 yes looks like collectd-libvirt should been a subpackage 2014-10-20 06:58:24 hi 2014-10-20 07:25:57 Hi 2014-10-20 07:26:18 I am compiling sdlmame and gives this error: http://pastebin.com/raw.php?i=un116zXC 2014-10-20 07:26:37 does anybody knows? 2014-10-20 07:58:51 AmatCoder: have you tried including stdio.h/cstdio? 2014-10-20 08:00:26 I tried with stdio.h 2014-10-20 08:00:52 the problem is that they want to get the typeinfo for an incomplete type 2014-10-20 08:01:21 ah, I read the wrong line 2014-10-20 08:01:24 and in musl FILE is not a complete object, you are only supposed to use FILE* 2014-10-20 08:03:28 at least i think that's the problem but i'm not c++ expert (and dont know how the dynamic_cast magic works) 2014-10-20 08:04:39 hmm, yes, FILE struct seems the problem here 2014-10-20 08:50:08 ncopa: hi, about libintl in kdelibs, i did some research and found that 2014-10-20 08:50:25 patch about libintl just set intl flag to linker 2014-10-20 08:51:09 because without it variable LIBINTL_LIBRARIES is empty and there is no intl flag in link command 2014-10-20 12:37:59 thx! 2014-10-20 12:39:43 jomat: the checksum failed but i assuem the patch was correct and apkbuild wrong 2014-10-20 12:43:35 http://sprunge.us/HPIK 2014-10-20 12:44:19 Something with whitespaces 2014-10-20 12:45:16 i just ran abuild checksum 2014-10-20 12:45:21 No Idea, glad it worked, will send github-urls in the future :-) 2014-10-20 12:51:31 ncopa: I think 1409238136-turnsole-25439@jmt.gr got lost somewhere, this is the patch: https://github.com/jomat/aports/commit/daa68538635027b8f4d2175c6f07dc1a3bcd31aa.patch 2014-10-20 12:53:25 hm, seems so yes. sorry 2014-10-20 12:53:30 np :-) 2014-10-20 13:54:51 fcolista2, liburcu == userspace-rcu ? 2014-10-20 13:55:13 that'e the name of the package 2014-10-20 13:55:14 we have it in main already 2014-10-20 13:55:20 with the name 'userspace-rcu' 2014-10-20 13:55:21 oh_ 2014-10-20 13:55:31 bah 2014-10-20 13:55:40 i didn't notice that 2014-10-20 13:55:49 umh 2014-10-20 13:56:35 it needs an update though 2014-10-20 13:56:56 and please do full configure line with --host, etc. 2014-10-20 13:57:16 it was needed for knot actually 2014-10-20 13:57:30 yes 2014-10-20 13:57:33 gdnsd etc. use it too 2014-10-20 13:58:59 one sec 2014-10-20 14:01:52 (phone) 2014-10-20 14:18:54 ok, back 2014-10-20 14:19:12 fabled, i'm gonna add host and build options on configure 2014-10-20 14:19:16 for knot 2014-10-20 14:19:24 question: 2014-10-20 14:19:53 how can know compile if userspace-rcu is in main ? 2014-10-20 14:20:28 btw i was not aware that liburcu == userspace-rcu 2014-10-20 14:20:56 ACTION jsut check with apk search liburcu  2014-10-20 14:21:17 ACTION just checked dumbly with apk search urcu  2014-10-20 14:22:27 fcolista2, no special action needed 2014-10-20 14:22:31 just depend on userspace-rcu 2014-10-20 14:22:37 testing stuff uses main automatically 2014-10-20 14:22:58 ok 2014-10-20 14:23:06 so i'll remove liburcu 2014-10-20 14:23:15 upgrade userspace-rcu 2014-10-20 14:23:26 and fix the configure options on knot 2014-10-20 14:24:45 right. thanks. 2014-10-20 14:25:50 thx for pointing it fabled. Btw, how can i avoid in the future this mistakes? I'm referring to userspace-rcu and liburcu 2014-10-20 14:25:59 mmm 2014-10-20 14:26:12 maybe read what kind of libs they are...but naming sometimes is misleading. 2014-10-20 14:26:14 "apk search urcu" did find userspace-rcu for me 2014-10-20 14:26:32 yes, i search for liburcu 2014-10-20 14:26:51 that finds it too 2014-10-20 14:27:05 not here 2014-10-20 14:27:06 you must've misspelled it 2014-10-20 14:27:14 maybe.. 2014-10-20 14:27:39 but searching by the '.so' library file name should find it via the provides 2014-10-20 14:28:02 maybe you are on old alpine? 2014-10-20 14:28:31 it's edge 2014-10-20 14:28:57 but apk update and apk search liburcu now return userspace-rcu 2014-10-20 14:29:10 ? 2014-10-20 14:29:30 huh 2014-10-21 08:38:38 hi 2014-10-21 08:41:43 Why can I not change the meta data of my own issues? 2014-10-21 08:42:09 made some mistakes in redmine (bugs.al) and I can not correct them 2014-10-21 08:44:08 what is your login name? 2014-10-21 09:10:17 thanks to francesco! 2014-10-21 09:17:12 having #3465 would be cool :-) 2014-10-21 09:18:15 we have haveged already 2014-10-21 09:21:53 but haveged can not use rdrand or hardware rng 2014-10-21 09:23:16 fair point 2014-10-21 09:33:56 I wanted to have an entropykey 2014-10-21 09:34:42 but someone build truerng with a big trustworthy american flag on it :) http://ubld.it/products/truerng-hardware-random-number-generator/ 2014-10-21 09:38:23 lol 2014-10-21 10:27:11 haveged is an extension to all other entropy sources you have. 2014-10-21 10:52:18 I have so many package requests :-* 2014-10-21 10:54:20 <[xming]> make them yourself :D 2014-10-21 10:56:04 <[xming]> I am still trying to get seafile working, but it doesn't with django 1.7, so I created a django15 pkg, is that the proper way? 2014-10-21 17:37:54 ncopa, ping 2014-10-21 17:38:02 ncopa, musl doesn't supply queue.h 2014-10-21 17:38:13 sys/queue.h that is 2014-10-21 17:38:27 but dhcpd needs it, does alpine supply its own queue.h? 2014-10-21 18:03:49 yes 2014-10-21 18:03:56 in libc-dev 2014-10-22 03:05:00 what's the best way to make a slightly customized version of an existing package? 2014-10-22 03:06:58 and how should I manage updates from the upstream package? 2014-10-22 06:22:48 shodan45: depends a bit on what the change is 2014-10-22 06:23:05 but i think it should be relatively easy to git clone the aports 2014-10-22 06:23:27 and then git pull get updates from upstream package 2014-10-22 08:04:09 clandmeter: you about? 2014-10-22 08:04:37 (morning all) 2014-10-22 08:04:58 <[xming]> hiya 2014-10-22 08:31:14 ScrumpyJack: yes i am 2014-10-22 08:48:13 i guess you are not? :) 2014-10-22 08:52:18 <[xming]> I am buidling seafile which only works with django 1.5, is it appropriate to create a pkg called django15? 2014-10-22 09:02:02 I'd prefer a dot 2014-10-22 09:02:05 django1.5 2014-10-22 09:02:15 this is what we were doing so far with various libraries 2014-10-22 09:03:02 autoconf2.13, goffice0.8, wxgtk2.8 2014-10-22 09:17:57 <[xming]> okay dot it is 2014-10-22 09:23:32 <[xming]> py-chardet (seafile's dep) -> http://sprunge.us/HBWT 2014-10-22 10:19:31 <[xming]> py-djanjo1.5 (seafile's dep) -> http://sprunge.us/cKWh 2014-10-22 10:36:25 <[xming]> py-django-pipeline (seafile's dep) -> http://sprunge.us/hLTh 2014-10-22 11:17:47 <[xming]> py-django-registation (seafile's dep) -> http://sprunge.us/iUbO 2014-10-22 11:25:06 <[xming]> py-django-simple-captcha (seafile's dep) -> http://sprunge.us/UJgH 2014-10-22 11:30:13 <[xming]> py-gunicorn (seafile's dep) -> http://sprunge.us/KNGj 2014-10-22 11:40:30 <[xming]> py-django-djblets (seafile's dep) -> http://sprunge.us/aVUP 2014-10-22 11:41:16 <[xming]> phew all the python stuff 2014-10-22 11:44:42 <[xming]> ccnet (seafile's dep) -> http://sprunge.us/jMgf 2014-10-22 11:45:54 [xming]: maybe push to a git repo we can pull from? 2014-10-22 11:46:55 <[xming]> ncopa, okay I will set that up 2014-10-22 11:47:15 github should be simple 2014-10-22 11:47:19 we have a mirror there 2014-10-22 11:47:34 https://github.com/alpinelinux/aports 2014-10-22 11:47:58 <[xming]> okay 2014-10-22 11:48:29 <[xming]> then I send PR through github? 2014-10-22 11:49:04 either that or in here say: please pull 2014-10-22 11:49:18 you would be the first to send us a PR :) 2014-10-22 11:49:24 PR from github should work 2014-10-22 11:49:31 would be cool to test it 2014-10-22 11:49:36 <[xming]> cool 2014-10-22 11:51:23 <[xming]> since seafile needs django1.5, I haven't made django-xyz to be dependant of django (either the 1.7 or 1.5) so they should work for both versions 2014-10-22 11:51:47 <[xming]> does alpine has something like virtual/* in Gentoo? 2014-10-22 19:57:00 our first github PR merged :) 2014-10-22 20:20:04 clandmeter: Nice, you guys have a pretty awesome distro 2014-10-22 20:20:13 clandmeter: by pressing a button? :) 2014-10-22 20:20:24 I was trying to port opensmtpd over but it needed way more work than I thought it would ;P 2014-10-22 20:20:40 Lots of glibc-isms 2014-10-22 20:21:02 barthalion: nah, reviewing first before commit. 2014-10-22 20:21:02 that's weird, I'd expect quite portable code from OpenBSD code 2014-10-22 20:21:10 s/code$/guys/ 2014-10-22 20:21:20 barthalion: I thought so too 2014-10-22 20:21:52 one would presume 2014-10-22 20:22:01 what did you run into? 2014-10-22 20:22:39 clandmeter: Just a bunch of undefined macros like SCOPE_DELIMITER 2014-10-22 20:23:28 And then I had to grab fts-dev 2014-10-22 20:23:37 But that's not really a portability issue 2014-10-22 20:24:49 Some stuff in sys/stat.h that was weirdly not defined (S_IXUSR and friends) 2014-10-22 20:25:42 clandmeter: what I'm asking is if it's enough to click "merge pull request" to make the magic happen 2014-10-22 20:26:17 barthalion: not really 2014-10-22 20:26:32 would only be possible if github is master 2014-10-22 20:26:44 I see 2014-10-22 20:26:58 but it aint that hard to do it by hand 2014-10-22 20:27:04 and you need to review it anyways 2014-10-22 20:27:08 if it builds and such 2014-10-22 20:27:21 yeah, sure thing 2014-10-22 20:27:31 the adv is, lots of users know github. 2014-10-22 20:27:48 and you can comment the code online if needed. 2014-10-22 20:28:09 that's a shame IMHO, I really like development mailing lists :) 2014-10-22 20:28:26 pull requests are completely not git-alike 2014-10-22 20:30:48 if we can provide whatever users like, thats best imho. 2014-10-22 20:31:51 you can send them to ML, you can sprunge them here (or advert your repo) or you can create a PR. 2014-10-23 07:20:14 <[xming]> hi I have a question about dropbear's dbclient. how do I add a passphrase to the private key? 2014-10-23 07:39:49 <[xming]> it seems that I am not the 1st one to do a PR on github :( 2014-10-23 07:39:52 <[xming]> https://github.com/alpinelinux/aports/pull/2 2014-10-23 07:40:11 :) 2014-10-23 07:40:33 few hours too late to be the first. 2014-10-23 07:40:54 <[xming]> meh 2014-10-23 07:41:10 <[xming]> who was the first? 2014-10-23 07:41:19 z3bra 2014-10-23 07:41:36 <[xming]> ah I see 2014-10-23 07:41:38 <[xming]> z3bratabs 2014-10-23 08:16:06 morning 2014-10-23 08:16:43 clandmeter: apologies, got distracted yesterday 2014-10-23 08:18:11 ScrumpyJack: morning 2014-10-23 08:20:04 [xming]: please make sure it builds after you modify apkbuild. 2014-10-23 08:20:19 rest looks good! 2014-10-23 08:21:07 i have a question about abuild and sub packages 2014-10-23 08:21:20 and dependancues :) 2014-10-23 08:21:28 and spelling :) 2014-10-23 08:21:32 lol 2014-10-23 08:21:50 <[xming]> clandmeter: sorry forgot about that one 2014-10-23 08:22:18 <[xming]> basically you have all the questions about subpkgs :p 2014-10-23 08:22:45 ScrumpyJack: i dont think we can help you with spelling. 2014-10-23 08:22:58 there should be better channels for such things. 2014-10-23 08:23:08 <[xming]> abuild should integrate aspell 2014-10-23 08:23:15 :) how can I be sure that abuild has worked out the dependancies of subpackages? 2014-10-23 08:23:40 <[xming]> by using the correct speeling of dependencies 2014-10-23 08:24:11 I have a bunch of subpackages, like libs, that are needed by the main package 2014-10-23 08:25:32 how can I make sure that my new_package_libs gets installed when new_package is installed? is it enough to put it in depends even though the package doesnt exist yet? 2014-10-23 08:28:08 i think abuild should automatically add the libs as depends 2014-10-23 08:28:42 <[xming]> if it contains ELF then it will be automatically, if not then just add it to depends 2014-10-23 08:29:40 <[xming]> ScrumpyJack: see this http://git.alpinelinux.org/cgit/aports/tree/testing/minetest/APKBUILD 2014-10-23 08:33:29 i guess he questions how abuild can scan for deps, when the pacakge is not yet ready. 2014-10-23 09:09:26 also, what's the prefered method of testing the dependencies when installing a new package (not in remote repo). create a private local repo? 2014-10-23 09:09:57 i think apk index can do that, but I've not tried 2014-10-23 09:11:54 <[xming]> isx the buildbot on vacation? 2014-10-23 09:12:33 <[xming]> damn I looked inthe wrong place again 2014-10-23 09:25:31 nice to see https://github.com/alpinelinux 2014-10-23 09:25:48 why cannot this be the main repo ? 2014-10-23 09:26:25 its possible 2014-10-23 09:26:30 it says "Readonly mirror of aports" may be "Official mirror of Alpinelinux aports" 2014-10-23 09:26:34 but everybody will have to push to github 2014-10-23 09:26:46 its already advertising PR 2014-10-23 09:26:54 there can be only one master repo 2014-10-23 09:27:01 and still saying "readonly" 2014-10-23 09:27:14 i think the wording is wrong 2014-10-23 09:27:22 well technically its correct 2014-10-23 09:27:35 <[xming]> it's correct IMHO 2014-10-23 09:27:39 s/well/while 2014-10-23 09:27:44 <[xming]> PR aren't pulled directly 2014-10-23 09:27:52 I know 2014-10-23 09:28:02 its via AL aports 2014-10-23 09:28:19 via git.a.o 2014-10-23 09:28:31 yes 2014-10-23 09:28:40 i had a discussion with ncopa about github 2014-10-23 09:28:58 <[xming]> it maybe a bit confusing, but it was clear to me 2014-10-23 09:29:27 ok, then 2014-10-23 09:29:32 <[xming]> clandmeter: BTW can you close my PR on github please? 2014-10-23 09:29:38 to switch, but we think some people would prefer not to have a github acc (ones that already have push access) 2014-10-23 09:31:54 [xming]: its still open? 2014-10-23 09:33:09 <[xming]> clandmeter: you've merge my seafile branch, but the PR on github is still open 2014-10-23 09:33:30 i had to rebase, i think thats the issue 2014-10-23 09:33:31 <[xming]> go git-wise it's done but github-wise it's still open 2014-10-23 09:33:36 what pkg uses testing/libsyncml ? 2014-10-23 09:33:53 grep aports 2014-10-23 09:35:22 found http://downloads.syncevolution.org/syncevolution/sources/ 2014-10-23 09:36:27 i would really prefer not to switch to github atm 2014-10-23 09:36:36 ok 2014-10-23 09:36:42 it would require developers sign up there (not big deal but...) 2014-10-23 09:36:47 and it would break the builders 2014-10-23 09:36:51 i would need reconfigure all builders 2014-10-23 09:37:10 its simply not worth it 2014-10-23 09:37:13 cost/benefit 2014-10-23 09:37:14 but iirc there is a mqtt pub 2014-10-23 09:37:22 i know 2014-10-23 09:37:30 ok 2014-10-23 09:38:12 ncopa: does algitbot output build failures? 2014-10-23 09:38:30 i think it does on stable branches 2014-10-23 09:38:32 but not edge 2014-10-23 09:38:36 <[xming]> github doesn't show more than 1000 files/dirs, that's a limitation 2014-10-23 09:38:37 can you enable that? 2014-10-23 09:38:57 when i push many commmits from smeobody else, its troublesome to verify them all. 2014-10-23 09:39:09 clandmeter: i can, but that also means it will stall the edge builders 2014-10-23 09:39:16 ...til someone cleaned up all testing packages 2014-10-23 09:39:40 it cant be done like now? 2014-10-23 09:39:51 im sure you can ;-) 2014-10-23 09:40:18 <[xming]> anyone has any idea how to add pssphrase to a priv key with dropbear? 2014-10-23 09:40:44 in main dhcpcd-ui failed 2014-10-23 09:40:47 and wireshark 2014-10-23 09:42:49 clandmeter: enabled halt on errors on build-edge-x86_64 2014-10-23 09:43:28 ncopa: why must hald be enabled? 2014-10-23 09:44:18 because buildrepo does not return with error otherwise 2014-10-23 09:45:18 the script checks: was building the repo success or not? 2014-10-23 09:45:25 ok, well halt is better then nothing. 2014-10-23 09:45:30 if something failed: oh it failed! 2014-10-23 09:45:43 but if you tell it to ignore errors and just continue, then 2014-10-23 09:45:50 well, is it a failure then? 2014-10-23 09:46:20 i'd need to refactor things 2014-10-23 09:46:36 an option to just msg, would be nice. 2014-10-23 09:46:48 and either add support for mqtt directly in buildrepo or abuild to do what we want 2014-10-23 09:47:35 i think it is good to fail early 2014-10-23 09:47:45 just have not had time to clean up the repos 2014-10-23 09:53:44 ncopa: build-edge-x86_64: files from v141022-35-g287dda5 uploaded. how is this "from" constucted? 2014-10-23 09:53:52 i guess v141022 is the last tag? 2014-10-23 09:54:02 man git describe 2014-10-23 09:54:05 yes 2014-10-23 09:54:28 i think it is: -- 2014-10-23 09:55:05 http://git-scm.com/docs/git-describe 2014-10-23 09:55:09 reading 2014-10-23 09:55:32 i couldnt find that hash 2014-10-23 09:56:36 ...-g 2014-10-23 09:56:47 287dda5 2014-10-23 09:57:31 "The hash suffix is "-g" + 7-char abbreviation for the tip commit of parent" 2014-10-23 09:57:34 yeah i added the g :) 2014-10-23 09:57:40 so i couldnt find it. 2014-10-23 09:59:14 <[xming]> seafile must be one of the most disturbing software I am going to run 2014-10-23 09:59:39 <[xming]> tone of dependencies (some out dated) and weird way of running it 2014-10-23 09:59:48 :-/ 2014-10-23 09:59:58 i thought seafile looked interesting 2014-10-23 10:00:06 <[xming]> but it's the best "cloud" I could find which meets my needs 2014-10-23 10:00:07 and might want set it up myself 2014-10-23 10:00:39 <[xming]> ncopa: I am almost there, had a trial run running 2014-10-23 10:01:00 <[xming]> still needs a lot of tweaking after install with my aport 2014-10-23 10:01:26 <[xming]> ncopa: at least all dependencies are in the aport now :D 2014-10-23 10:01:36 this is interesting :) http://reverbrain.com/ 2014-10-23 10:01:37 i also saw some syncing tool that used torrents 2014-10-23 10:01:47 <[xming]> seafile has free android and ios client, which is a big plus 2014-10-23 10:01:51 you should try syncthing 2014-10-23 10:01:56 i added it to aports recently 2014-10-23 10:02:02 syncthing was what i saw 2014-10-23 10:02:08 looks interesting indeed 2014-10-23 10:02:09 its not based on torrents 2014-10-23 10:02:46 hm 2014-10-23 10:02:47 ok 2014-10-23 10:02:49 <[xming]> itn'st it a btsync clone? 2014-10-23 10:02:56 no 2014-10-23 10:03:02 In episode 456 of SecurityNow!, 20 May 2014, host Steve Gibson praised Syncthing as a potential open-source replacement for BitTorrent Sync.[4] 2014-10-23 10:03:27 a replacement isnt a clone :) 2014-10-23 10:03:29 i also thought it was using bt 2014-10-23 10:03:32 yeah 2014-10-23 10:03:51 it will connect to other peers by itself 2014-10-23 10:04:00 it just needs a resolver 2014-10-23 10:04:11 which you cna install yourself, or use the public one. 2014-10-23 10:04:18 <[xming]> does it have a free ios app? 2014-10-23 10:04:43 sorry i dont have any apple device. 2014-10-23 10:04:50 but it does have one for android 2014-10-23 10:05:10 and its free 2014-10-23 10:05:14 seems android only 2014-10-23 10:05:14 <[xming]> https://github.com/syncthing/syncthing/issues/102 2014-10-23 10:05:27 it has just rebranded 2014-10-23 10:05:33 <[xming]> oh it's written in Go 2014-10-23 10:05:50 <[xming]> hmm should I drop seafile? hmm tempting, just need ios client 2014-10-23 10:06:15 i think seafile looks like a nice alternative 2014-10-23 10:06:19 its called pulse now. 2014-10-23 10:07:27 seafile looks a bit too commercial to me. 2014-10-23 10:07:35 <[xming]> seafile has free ios client 2014-10-23 10:07:37 <[xming]> but 2014-10-23 10:07:47 can you sync calendar/contacts with pulse/syncthing? 2014-10-23 10:07:56 <[xming]> it's a OSS <-> commercial conflict of interest 2014-10-23 10:08:18 <[xming]> too much dependencies, hard to maintain, maybe even hard to upgrade, ... 2014-10-23 10:09:11 <[xming]> I use zimbra, and zpush (I have instlled it here yet) can sync cal/contacts via active sync 2014-10-23 10:10:40 <[xming]> my kid just wanted some picture on my android phone on the ipod, it was damn hard to do that w/o cords 2014-10-23 10:10:51 <[xming]> seafile solves that for me 2014-10-23 10:15:43 i dont think it will take long for an ios client for syncthing. there is lots of developemtn going on there. 2014-10-23 10:17:53 <[xming]> Go doesn't work on ios and Go doesn't do dyn linking right now 2014-10-23 10:18:57 <[xming]> but idd synthing looks a lot better for the futhur, and seafile solves my problem now 2014-10-23 10:19:11 <[xming]> s/futher/future 2014-10-23 10:22:17 well, it does, but you need to use gcc-go for that 2014-10-23 10:31:06 <[xming]> gcc-go never worked well here, seg faulting all over the place 2014-10-23 11:17:31 so, what's the best way to check dependencies between aports not in a repo? create a local repo? is this done with apk index? 2014-10-23 12:59:24 rnalrd: thanks! 2014-10-23 12:59:27 doh 2014-10-23 12:59:37 it's tcpflow 2014-10-23 12:59:46 need it in 3.0-stable 2014-10-23 12:59:48 backporting 2014-10-23 13:01:27 godo 2014-10-23 13:01:30 good 2014-10-23 13:01:44 i have only xen fix and then we are ready for 3.0.6 2014-10-23 13:27:43 ncopa: i noticed python is still in alpine-xen 3.1.0 Is that deliberate? 2014-10-23 13:28:48 3.1.0 not released? 2014-10-23 13:29:26 try those: http://dev.alpinelinux.org/~ncopa/xen/ 2014-10-23 13:56:22 yes, from there. 2014-10-23 13:56:39 I have 141014 2014-10-23 13:57:36 it still has xend and python :( 2014-10-23 14:32:49 YX9KwoR6rk2l: ^ ddate is back :-D 2014-10-23 14:34:32 so 3.0.6 is 3.1 alpha 2? 2014-10-23 14:46:59 3.0.6 is stable branch 3.0 2014-10-23 14:47:24 3.1 (alpha) is currently the same as edge 2014-10-23 14:47:47 http://wiki.alpinelinux.org/wiki/Alpine_Linux:Releases 2014-10-23 14:50:14 ah, ok 2014-10-23 14:51:34 ncopa: what's the best way to test dependencies in packages that are not in the public repos? 2014-10-23 14:51:43 create a local repository? 2014-10-23 14:51:58 using apk index? 2014-10-23 14:52:01 test dependencies? 2014-10-23 14:52:02 yes 2014-10-23 14:52:14 i think so 2014-10-23 14:52:31 yes, test depensdencies between packages that are on my local dev box 2014-10-23 14:52:49 local repo should be the easist way to handle that 2014-10-23 14:54:45 is apk index enough? 2014-10-23 14:55:31 can you put a local fs in /etc/apk/repos ? 2014-10-23 14:56:30 jomat: \<3/ 2014-10-23 14:57:02 how can these people live without ddate? 2014-10-23 14:59:04 ScrumpyJack: i think you need sign it too, with abuild-sign APKINDEX.tar.gz 2014-10-23 15:00:58 checking for exiv2/image.hpp... no 2014-10-23 15:00:58 checking for git... yes 2014-10-23 15:00:58 configure: git commit v141022-39-gfef24a3-dirty 2014-10-23 15:01:02 bah 2014-10-23 15:01:09 tcpflow does stupid things 2014-10-23 15:19:07 ncopa: thanks, yes i wondered about signing 2014-10-23 15:19:26 excellent. there is no docs i can see on the wiki so I'll add 2014-10-23 15:19:49 "creating your own apk repository" ) 2014-10-23 15:22:34 how much data is transfered out of a typical alpine repo or mirror? 2014-10-23 15:56:19 not much I guess 2014-10-23 15:56:31 I mean, I host two mirrors, Arch and Alpine 2014-10-23 15:57:13 it's maybe 3-4GB daily for both 2014-10-23 15:58:31 and you can figure it out that more than 50% of that is for Arch 2014-10-23 16:06:56 barthalion: hi, about your package mysql, why did your turn off embedded mysql ? 2014-10-23 16:07:16 s/your/you 2014-10-23 16:07:43 it's mine? 2014-10-23 16:07:48 I bump it from time to time 2014-10-23 16:07:52 but it's hardly mine 2014-10-23 16:08:11 please send us a patch that enables it and I'll merge it yet today :) 2014-10-23 16:09:04 ahhh, sorry no, it's not your :/ 2014-10-23 16:11:06 main/mysql: -> http://sprunge.us/EaVi 2014-10-23 16:22:14 k0r10n: thanks 2014-10-23 16:25:17 huh 2014-10-23 16:26:35 why did it fail when it should not be building at all? 2014-10-23 16:39:52 hmmm, strange apk update fails with wget getting a EPIPE - what is wget piping into? 2014-10-23 17:06:16 barthalion: thanks for the info. which alpine mirror is yours? 2014-10-23 17:12:27 ScrumpyJack: bpiotrowski.pl 2014-10-23 17:16:10 this? http://files.bpiotrowski.pl/archlinux/ 2014-10-23 17:18:09 is it one of these? http://nl.alpinelinux.org/alpine/MIRRORS.txt 2014-10-23 17:19:30 do we even use MIRRORS.txt for anything? 2014-10-23 17:19:45 the official list is held in main/alpine-mirrors 2014-10-23 17:20:32 MIRRORS.txt is missing at least two addresses… 2014-10-23 17:20:35 mirror.bpiotrowski.pl 2014-10-23 17:25:20 lua gen-txt.lua < mirrors.yaml > MIRRORS.txt 2014-10-23 17:30:24 yeah, http://nl.alpinelinux.org/alpine/MIRRORS.txt is out of date if you look at the alpine-mirrors apk 2014-10-23 17:30:49 perhaps http://nl.alpinelinux.org/alpine/MIRRORS.txt should be autogenerated 2014-10-23 17:30:58 ncopa: ↑ 2014-10-23 17:31:54 probably for the new site :) 2014-10-23 18:46:48 ScrumpyJack: Do you know http://wiki.alpinelinux.org/wiki/Alpine_Linux_package_management#Packages_and_Repositories and http://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package ? 2014-10-23 18:48:27 I've learned from those pages that my packages are signed, stored and indexed in ~/packages/ 2014-10-23 18:49:32 ScrumpyJack: I added '@local /home/jomat/packages/testing' to my /etc/apk/repositories 2014-10-23 18:50:38 ScrumpyJack: when I create dependencies for packets in my local aports-tree, I install them with 'apk add somepkg@local' 2014-10-23 18:50:58 perhaps that helps :-) 2014-10-23 19:15:02 that's great, thanks, sounds like it's what I need 2014-10-23 19:16:13 do you run apk index? 2014-10-23 19:16:31 i mean, do you run apk index on ~/packages/ ? 2014-10-23 19:18:20 no 2014-10-23 19:18:57 just abuild -r and if that's successful apk add pkg@local 2014-10-23 19:19:12 hmm 2014-10-23 19:19:20 ehh... apk update 2014-10-23 19:19:24 between them 2014-10-23 19:19:27 i think 2014-10-23 19:20:00 i would like to create a private repo where I can point users to test packages that they request. I'll look into it 2014-10-23 19:20:12 any tips welcome of course :) 2014-10-23 19:24:08 Then the other users need your public key installed in /etc/apk/keys/ 2014-10-23 19:34:40 sounds about right :) 2014-10-23 19:35:20 there is probably an option to ignore the keys altogether, i'm new still new to alpine linux 2014-10-24 07:20:29 morning 2014-10-24 10:28:39 oops 2014-10-24 10:31:12 ncopa: where these errors come from? you bumped something in perl that broken almost everything? 2014-10-24 10:38:04 i upgraded perl 5.20 some time ago 2014-10-24 10:38:09 but didnt rebuild testing 2014-10-24 11:51:57 <[xming]> how do I do `realpath -s` w/o installing coreutils? 2014-10-25 05:30:43 morning 2014-10-25 05:31:58 I have been reading the discussion about nix package manager from irclogger... 2014-10-25 05:32:36 ... and I have to say that I use it to run glibc programs on Alpine 2014-10-25 05:32:45 like chromium 2014-10-25 05:34:29 no need to patch to have PaX flags 2014-10-25 05:35:08 packages from nixpkgs are 'grsec-aware' ;) 2014-10-25 05:39:14 oops, I wrote it in the wrong channel...sorry 2014-10-25 08:51:07 still interesting :) 2014-10-25 09:22:05 barthalion: me? 2014-10-25 09:43:36 yeah 2014-10-25 09:50:28 well, I can tell you as I use nix 2014-10-25 09:50:46 keep in mind that it has some disadvantages 2014-10-25 09:52:26 mainly it is a bloated system 2014-10-25 09:52:46 because they does not split packages 2014-10-25 09:53:45 so if one package needs some lib_gcc as dependency, it install the whole gcc suite 2014-10-25 09:55:11 yes, I have read about it 2014-10-25 09:55:26 I need to find some time soon to finally play with nix 2014-10-25 09:56:29 meeeh. that sounds aweful 2014-10-25 09:56:48 quite the opposite of lean/mean alpine. which fits into a few megs. 2014-10-25 09:57:59 well, it is the only way to run some apps incomplatible with musl 2014-10-25 09:58:08 without chroot 2014-10-25 09:58:17 definitely there are other alternatives 2014-10-25 09:58:21 like porting to musl 2014-10-25 10:01:25 well, not the only but the easier then :) 2014-10-25 10:02:31 easier for whom? 2014-10-25 10:02:46 shortterm/longterm? 2014-10-25 10:02:48 etc 2014-10-25 10:03:03 for me at least 2014-10-25 10:05:57 Use of nix can be considered as temporary until musl fix those rough corners 2014-10-25 10:06:07 s/fix/fixes 2014-10-25 10:06:58 maybe it's not musl that needs fixing, but the other programs implementing stupid gnuisms, or their own mallocs and other crap? 2014-10-25 10:07:26 musl it seems to me points out the fuckups in widely used sw. 2014-10-25 10:09:58 I'm not criticizing musl, I merely point out an alternative... 2014-10-25 10:10:55 ...until either musl or software can be used. 2014-10-25 10:12:18 the last sentence has no sense :( 2014-10-25 10:12:33 hehe 2014-10-25 10:12:42 my english is not good enough 2014-10-25 10:14:06 dontworry ;) 2014-10-25 10:14:40 ;) 2014-10-25 12:10:20 morning 2014-10-25 12:10:58 anyone tried testing/virtualbox-additions-grsec? 2014-10-25 16:50:20 algitbot: liar 2014-10-25 21:56:21 how can I create an alpine rootfs for armhf ? 2014-10-25 21:56:47 apk add --root /path/to/root --arch armhf ends in ERROR: Unable to lock database: No such file or directory 2014-10-25 21:56:54 and ERROR: Failed to open apk database: No such file or directory 2014-10-25 22:06:39 *PING* 2014-10-25 22:20:02 ok got it ... doku is lagging a lot of this topic 2014-10-26 12:21:44 :| 2014-10-26 12:22:20 fun 2014-10-26 13:03:08 clandmeter, around ? 2014-10-26 13:04:34 any corosync alternative ? 2014-10-26 13:05:29 vkrishn: i dont know 2014-10-26 13:05:44 i think it needs some love. 2014-10-26 13:05:49 which i dont have at this moment. 2014-10-26 13:06:00 noticed it moved to unmaintained 2014-10-26 13:06:03 atleast not for corosync :) 2014-10-26 13:06:15 yes, it doesnt build 2014-10-26 13:06:18 and hasnt for some time. 2014-10-26 13:06:36 bcoz, libqb needs to be moved to testing 2014-10-26 13:06:50 which I think is in unmaintained 2014-10-26 13:06:57 yes that was it. 2014-10-26 13:07:02 corosync requires libqb 2014-10-26 13:07:16 dont know who moved libqb and what the status is. 2014-10-26 13:07:28 Uggedal 2014-10-26 13:07:31 :) 2014-10-26 13:07:53 commit 923bc4ad78801603 2014-10-26 13:07:58 maybe for good reasons 2014-10-26 13:08:09 I thought so, is he around ? 2014-10-26 13:08:43 i moved on 2014-10-26 13:08:45 err 2014-10-26 13:08:47 he moved on 2014-10-26 13:09:13 [*] 2014-10-26 13:10:09 if you check his github page, you probably know where. 2014-10-26 13:12:27 it wasnt such a good idea to let builders stop on error. 2014-10-26 13:12:35 nothing gets synced now. 2014-10-26 13:13:01 speaking of which, can we finally move everything older than a month from testing to /dev/null? 2014-10-26 13:13:33 the probably need to move testing/pacemaker too 2014-10-26 13:14:29 and check main/ocfs2-tools/build.patch 2014-10-26 13:18:02 and main/drbd has config option --with-pacemaker, don't know if it means usage of testing/pacemaker 2014-10-26 13:19:35 barthalion: i think its better to discuss that on ML 2014-10-26 13:20:00 atleast not in the weekend on irc. 2014-10-26 13:21:46 sunday afternoon is the best moment to do big underdiscussed changes 2014-10-26 13:22:54 :) 2014-10-26 13:25:30 maybe update http://clusterlabs.github.io/libqb/ to see if it builds 2014-10-26 13:32:34 http://www.irclogger.com/.alpine-devel/2014-06-02#1401752091 2014-10-26 13:32:56 thanks algibot 2014-10-27 08:30:07 ncopa: builders dont upload pkgs? 2014-10-27 08:44:52 clandmeter: not til we have cleaned up the testing repo 2014-10-27 08:45:37 i do as fast as i can 2014-10-27 08:45:45 feel free to help though 2014-10-27 08:46:29 thos are knonw to be broke: http://sprunge.us/BAPS 2014-10-27 08:46:57 what i have done, try upgrade to latest version, if it fails move to unmaintained 2014-10-27 08:50:29 ncopa: ive been doing that this weekend. seems like a neverendign list 2014-10-27 08:51:19 :-/ 2014-10-27 08:51:30 but i think we need to do it at some point 2014-10-27 08:51:32 but ill try to help 2014-10-27 08:51:36 thanks 2014-10-27 08:51:36 <[xming]> morning 2014-10-27 08:51:45 actually 2014-10-27 08:51:52 the list was shorter than expected 2014-10-27 08:52:31 so good job! 2014-10-27 08:53:12 <[xming]> I have basically the seafile +seahub aports ready, but I am not sure for certain things, so I am not going to make a PR 2014-10-27 08:53:54 <[xming]> basically those two things are so interwinded, it's better to make one package of it, but on the other hand seafile cn be used alone 2014-10-27 08:54:17 <[xming]> and there comes all the dependencies which I am not sure which package should have it 2014-10-27 08:55:53 <[xming]> ncopa: if you are still interested in seafile, I can give you the aports in the current state, maybe you have a better idea of how it should be done? 2014-10-27 09:05:36 ncopa, could this be checked, https://dpaste.de/thJL/raw, I think some may be needed or being used/stable 2014-10-27 09:07:15 why is zoneminder in that list? 2014-10-27 09:07:20 its in main 2014-10-27 09:08:10 ncopa: gnome is a mess in main 2014-10-27 09:08:19 we should really clean that up. 2014-10-27 09:10:44 I could suggest folder /desktop to move them to (gnome+kde) , i.e heavier wm 2014-10-27 09:11:20 they would build like testing, but can be triggered to stall 2014-10-27 09:12:13 we dont have full gnome, just parts which are left overs from gnome 2 2014-10-27 09:12:47 and we have somebody who is sending patches for kde, so if that works it should stay in main. 2014-10-27 09:14:32 unlike /testing , /desktop moves along with /main in stable 2014-10-27 09:14:49 its just a start to make dir 2014-10-27 09:15:48 we've had it before, and ncopa decided to put it in main. I don't remember the reason. 2014-10-27 09:16:02 like /misc /etc 2014-10-27 09:16:12 ok 2014-10-27 09:17:29 , /nongpl 2014-10-27 09:36:50 hello all, does alpine have support for getcontext ? 2014-10-27 09:39:04 $ nm -D /lib/libc.musl-x86_64.so.1 | grep getcontext 2014-10-27 09:39:07 seems not 2014-10-27 09:39:12 :/ 2014-10-27 09:39:51 i think posix decided to deprecate it and let pthread do that job 2014-10-27 09:44:00 ok, ty 2014-10-28 09:34:04 Great news! Chromium was built successefully on 64 Alpine 2014-10-28 09:34:52 cool! 2014-10-28 09:35:21 normally chromium has its own allocator that overrides libc malloc etc 2014-10-28 09:35:28 i guess that was disabled 2014-10-28 09:35:31 yes 2014-10-28 09:35:46 firefox has that too, but with firefox you can disable it 2014-10-28 09:36:33 in chromium too, use_allocator=none 2014-10-28 09:37:50 ncopa: can you spin a list of failures? 2014-10-28 09:38:06 else we are doing the same thing twice 2014-10-28 09:38:12 ok 2014-10-28 09:38:13 i will 2014-10-28 09:39:17 maybe its also a little more motivating to see its shrinking... :) 2014-10-28 09:40:24 http://sprunge.us/bADO 2014-10-28 09:40:48 i can start from below 2014-10-28 09:41:21 ok 2014-10-28 09:45:07 is this APKBUILD for v8 acceptable -> http://sprunge.us/DWga ? 2014-10-28 09:50:11 k0r10n: i think you can remove the depends 2014-10-28 09:50:14 depends= 2014-10-28 09:50:27 abuidl should autodetect SO depends 2014-10-28 09:50:53 yes, sure, always forgot... 2014-10-28 09:51:03 this version of v8 need to build chromium 2014-10-28 09:51:35 they dont provide downloadable tarballs? 2014-10-28 09:52:08 they don't have such version tarball :/ 2014-10-28 09:52:22 and for chromium need spicific trunk 2014-10-28 09:53:16 link failed with other version of v8 2014-10-28 09:55:42 why bother packaging it independently? 2014-10-28 09:56:12 I see no clear advantages of that 2014-10-28 09:56:49 each chromium may want to use different v8 release, always out of date or modified somehow 2014-10-28 09:57:15 yes, each version use it's own v8 version 2014-10-28 09:57:50 barthalion: you mean packaging v8 independently or gyp and icu ? 2014-10-28 09:58:11 I'm talking about v8 2014-10-28 09:58:27 because segmentation fault 2014-10-28 09:58:30 I'd let it pull v8 on its own 2014-10-28 09:58:48 when it tries to run mksnapshot 2014-10-28 09:59:10 sure, but packaging v8 under the package name v8 you're polluting packages namespace 2014-10-28 09:59:14 current release of v8 is 3.29.88.10 2014-10-28 09:59:21 or even 3.30.something 2014-10-28 09:59:56 3.26.31 2014-10-28 10:00:19 I'm quite sure it's 3.30 2014-10-28 10:00:24 git tags don't lie 2014-10-28 10:00:36 u mean mainstream ? 2014-10-28 10:00:49 yes 2014-10-28 10:01:08 maybe 2014-10-28 10:01:14 k, have to go.. 2014-10-28 10:03:55 <[xming]> does alpine has soft depends? Like Debian's recommended? 2014-10-28 10:04:12 no, i dont think so 2014-10-28 10:04:26 how does those work? 2014-10-28 10:04:42 something like optdepends in Arch, yes? 2014-10-28 10:04:57 what does those do? 2014-10-28 10:04:59 packages that user might want to have, but doesn't have to 2014-10-28 10:05:13 how does package manager handle those? 2014-10-28 10:05:18 just give info about them? 2014-10-28 10:05:20 <[xming]> like seafile I am doing now 2014-10-28 10:05:28 yes, Arch just prints the possibility 2014-10-28 10:05:34 <[xming]> it doens't need seahub/python/.... to run 2014-10-28 10:05:34 Debian probably install them by default :D 2014-10-28 10:05:53 <[xming]> but seafile needs those to setup and some utils 2014-10-28 10:06:00 but it doesnt have any influence over the dependency tree? 2014-10-28 10:06:06 correct, sir 2014-10-28 10:06:09 <[xming]> IIRC Debian prints them by default 2014-10-28 10:06:21 so it is just a FYI... 2014-10-28 10:06:26 Debian has also suggested, maybe those are installed 2014-10-28 10:06:37 I'm happy avoiding Debian as much as I can 2014-10-28 10:06:54 <[xming]> maybe Debian users can turn to install those on by default 2014-10-28 10:07:41 <[xming]> seafile aports are basically ready, just there are things I don't know how to handle them 2014-10-28 10:08:05 <[xming]> like bundle seafile + seahub so everything is included, but bloated 2014-10-28 10:09:03 <[xming]> also seafile + seahub can be run in multiple instances, but some file + dirs need to copied (or hard linked, not dirs obviously) and how to handle those perms 2014-10-28 10:09:41 <[xming]> so currently I have 2 aports and cp + chmod in the post install script, which give me a yellow warning 2014-10-28 10:12:57 <[xming]> I like that upstart is moved to unmaintained :p 2014-10-28 10:13:09 yeah, i considered just wipe it... 2014-10-28 10:13:38 <[xming]> ncopa: any guidance how I should solve the issues with seafile? 2014-10-28 10:14:43 i havent looked closely at this specific case 2014-10-28 10:14:49 <[xming]> hoorray mode systemd discussions :/ http://linux.slashdot.org/story/14/10/28/0026200/debate-over-systemd-exposes-the-two-factions-tugging-at-modern-day-linux 2014-10-28 10:14:55 but i tend to generaly only install whats needed to run minimal 2014-10-28 10:15:07 and let user fix the rest 2014-10-28 10:15:16 <[xming]> okay 2014-10-28 10:15:22 <[xming]> I like that6 approach 2014-10-28 10:15:34 <[xming]> and about chmoding in the post-install? 2014-10-28 10:15:52 so if you have an app that has support for both mysql and postgres, and require one of those, i tend to install none 2014-10-28 10:15:58 and let user pick 2014-10-28 10:16:21 alpine linux tries stay out of the way 2014-10-28 10:16:38 chmodding in post-install? 2014-10-28 10:16:44 <[xming]> exactly what I want :D 2014-10-28 10:16:46 try avoid that 2014-10-28 10:17:02 why do you need chmod in post-install? 2014-10-28 10:17:09 <[xming]> seahub is the web front-end of seafile 2014-10-28 10:17:15 ah 2014-10-28 10:17:24 and the web server needs access 2014-10-28 10:17:29 <[xming]> it needs to have perm of the user running it 2014-10-28 10:17:32 write access or read access 2014-10-28 10:17:43 we have a groups for that 2014-10-28 10:17:46 wwwdata iirc 2014-10-28 10:17:55 the idea is that webservers are in that group 2014-10-28 10:17:56 <[xming]> I have n't figure the whole thing out yet, but at least write access on some parts 2014-10-28 10:18:06 <[xming]> it is a web server 2014-10-28 10:18:13 <[xming]> in pyton 2014-10-28 10:18:15 and you give that group write permissions to the needed dir 2014-10-28 10:18:36 then you create a user for it, and add that user to wwwdata group 2014-10-28 10:18:48 specially if it is possible to run with other webservers 2014-10-28 10:18:57 <[xming]> the idea is that a system can have multiple instances of, running under diff uid/gid 2014-10-28 10:19:08 <[xming]> on diff ports 2014-10-28 10:19:17 hm 2014-10-28 10:19:26 which is configurable 2014-10-28 10:19:27 <[xming]> either as webserver or as fgci server 2014-10-28 10:19:38 <[xming]> yes configurable in the confd 2014-10-28 10:19:56 <[xming]> but during the installation of the pkg, I want to make a default instance 2014-10-28 10:20:04 i think we can provide a usable default 2014-10-28 10:20:14 <[xming]> which will be for simple installs or serve as example 2014-10-28 10:20:31 <[xming]> and I have to copy those file and chmod them 2014-10-28 10:20:49 and if user edits conf.d with more instances as different users, then try get out of the way for user 2014-10-28 10:20:58 but user will have to chmod the dir himself 2014-10-28 10:21:15 optionally, you could do checkpath --directory --owner from init.d script 2014-10-28 10:21:33 <[xming]> when user adds more instances, he will need to edit confd and mkdir/cp/chmod himself, like the default instance 2014-10-28 10:22:10 <[xming]> yes but init.d is too late, as during the setup, it needs some write access already 2014-10-28 10:22:15 ok 2014-10-28 10:22:17 hm 2014-10-28 10:22:35 <[xming]> and it's better to do setup AS the uid which is going to run that instance 2014-10-28 10:22:48 difficult to give advice when i dont understand how it is though to work 2014-10-28 10:23:06 we could start simple, with support for a single instance 2014-10-28 10:23:21 <[xming]> then maybe I should just finish it like this now? and let users to give feedbacks? 2014-10-28 10:23:27 yes 2014-10-28 10:23:29 sounds good 2014-10-28 10:23:39 <[xming]> great thanks 2014-10-28 10:23:43 i will try set up one for myself at one point 2014-10-28 10:24:16 <[xming]> can I print stuffs in the termanl after the intallation of a pkg? 2014-10-28 10:24:38 <[xming]> just to inform users how to setup, and create an instance, ... 2014-10-28 10:24:54 <[xming]> Imean how can I print on the term 2014-10-28 10:34:24 <[xming]> I think this deserves a wiki page 2014-10-28 12:09:01 got to love these msg's configure: error: Could not find a version of the library! 2014-10-28 12:09:34 <[xming]> yeah, seafile + seahub aports are ready, got rid of chown in the post-install, I am pretty happy about them now, considering the upstream way of doing things 2014-10-28 12:09:42 <[xming]> https://github.com/alpinelinux/aports/pull/5 2014-10-28 12:10:52 <[xming]> do unmaintained aports get built? 2014-10-28 12:10:58 nope 2014-10-28 12:11:09 they are there for a reason. 2014-10-28 12:11:17 <[xming]> or are you guys moving unbuidables to unmaintained? 2014-10-28 12:11:19 mostly because they dont build 2014-10-28 12:11:33 depends a bit 2014-10-28 12:11:40 everthing can be build 2014-10-28 12:12:03 when its in testing, we asume nobody uses it. 2014-10-28 12:12:10 <[xming]> hmm, I need to remember to fish in the unmaintained when I need something :D 2014-10-28 12:12:25 yes, its better bigger and bigger 2014-10-28 12:12:57 <[xming]> all aports I have submitted are actively used by me, except the desktop counterparts 2014-10-28 12:14:26 <[xming]> is the buildbot now stuck when it encounters an error? 2014-10-28 12:15:32 <[xming]> why are we getting so much more failed spams? 2014-10-28 12:16:41 yes 2014-10-28 12:16:56 we never cleaned up testing 2014-10-28 12:17:22 <[xming]> that's hell of ungrateful job :/ 2014-10-28 12:17:23 so it was accepting failed build 2014-10-28 12:17:34 its freaking horrible 2014-10-28 12:17:44 but else ncopa has to do it. 2014-10-28 12:17:50 and he is much better at other thigns. 2014-10-28 12:19:11 freaking configure error is so unclear, i cant find the missing dep. 2014-10-28 12:19:32 <[xming]> which package? 2014-10-28 12:20:43 mkvtoolnix 2014-10-28 12:20:47 clandmeter: you want me to look at it? 2014-10-28 12:21:10 i had it build local 2014-10-28 12:21:17 http://sprunge.us/fchX 2014-10-28 12:21:17 had lots of crap in my tree 2014-10-28 12:21:19 updated list 2014-10-28 12:21:31 so the build server doesnt accept it 2014-10-28 12:21:38 ncopa: sure shoot 2014-10-28 12:21:43 its should be buildable. 2014-10-28 12:22:29 <[xming]> lemme try freeciv 2014-10-28 12:22:33 configure: error: Could not find a version of the library! 2014-10-28 12:22:34 heh 2014-10-28 12:22:43 useful error message :) 2014-10-28 12:23:07 i told you ;-) 2014-10-28 12:23:18 freeciv should be ok 2014-10-28 12:23:24 maybe i missied a patch 2014-10-28 12:23:29 yeah, some of them should build 2014-10-28 12:23:29 <[xming]> building and let's see 2014-10-28 12:23:34 i thinnk most of them are done 2014-10-28 12:23:35 i think the buidler change the order 2014-10-28 12:23:40 but im so freaking hasty 2014-10-28 12:23:41 so some of those are fixed already 2014-10-28 12:23:46 i keep forgetting patches :D 2014-10-28 12:24:21 exfat should be also ok 2014-10-28 12:24:33 "the library" 2014-10-28 12:24:35 is boost 2014-10-28 12:24:52 <[xming]> and I forget abuild checksum all the time 2014-10-28 12:25:08 ncopa: i thought it was boost. but i got boost-dev 2014-10-28 12:25:20 <[xming]> maybe ncopa can make abuild to checksum before building 2014-10-28 12:25:42 [xming]: dont start about that again... ;-) 2014-10-28 12:26:11 <[xming]> oh again? That must be before I started using alpine 2014-10-28 12:26:17 ncopa: who won that price eventually? 2014-10-28 12:26:25 i think it was myself :) 2014-10-28 12:26:29 <[xming]> or I just missed that 2014-10-28 12:26:44 nah its long ago 2014-10-28 12:26:53 more then a year maybe 2014-10-28 12:27:02 <[xming]> what's the prize? a free Alpine 3.0.6 iso? 2014-10-28 12:27:14 fix what you requested 2014-10-28 12:27:17 <[xming]> surely before my time 2014-10-28 12:27:35 the prize was to write a git push hook 2014-10-28 12:27:45 that catches missing checksums 2014-10-28 12:27:49 and rejects the git push 2014-10-28 12:28:05 and a forkbomb :) 2014-10-28 12:29:09 i think that list, all of them should be fixed. 2014-10-28 12:29:14 accept mkvtoolnix 2014-10-28 12:29:24 or i messed them up somehow 2014-10-28 12:29:30 with leftover deps 2014-10-28 12:30:16 ah 2014-10-28 12:30:19 i know why it fails 2014-10-28 12:30:26 stupid configure script 2014-10-28 12:30:37 for libextension in `ls $BOOSTLIBDIR/libboost_system*.{so,dylib,a}* 2>/dev/null | sed 's,.*/,,' | sed -e 's,^lib\(boost_system.*\)\.so.*$,\1,' -e 's,^lib\(boost_system.*\)\.dylib.*$,\1,' -e 's,^lib\(boost_system.*\)\.a.*$,\1,'`; do 2014-10-28 12:30:54 find one thing that does not work in busybox ash 2014-10-28 12:31:28 ls $BOOSTLIBDIR/libboost_system*.{so,dylib,a}* 2014-10-28 12:31:32 does not 2014-10-28 12:32:00 *.{so,dylib,a}* is bashism 2014-10-28 12:32:03 ah so its probably bash which was installed 2014-10-28 12:32:06 yup 2014-10-28 12:32:12 freaking shit 2014-10-28 12:32:21 took me 20 minutes of my life 2014-10-28 12:32:28 i want my money back 2014-10-28 12:32:32 :) 2014-10-28 12:32:59 if i would have gotten money for each time of those bash blah things... 2014-10-28 12:33:04 i'd been rich by now 2014-10-28 12:33:42 you are 2014-10-28 12:33:45 you have alpine 2014-10-28 12:34:15 the new ones are popping up 2014-10-28 12:34:15 more breakage? 2014-10-28 12:34:17 yes 2014-10-28 12:34:21 :| 2014-10-28 12:34:29 they were hidden due to deps was not built 2014-10-28 12:34:35 i dont think we have many left now 2014-10-28 12:35:23 pacemaker can be moved to unmaintained 2014-10-28 12:35:27 we did so with the rest of it. 2014-10-28 12:35:40 <[xming]> freeciv is golden 2014-10-28 12:36:09 [xming]: if you find stuff which works, just request it to be moved to main. 2014-10-28 12:36:18 else it will end in the gravejard :) 2014-10-28 12:36:46 <[xming]> I've done that with some of my own aports 2014-10-28 12:37:34 <[xming]> but e.g. minetest I only use the server part and never tried the client (still don't have alpine desktop here) can I also request it? 2014-10-28 12:39:55 request? 2014-10-28 12:40:30 isnt py-xen shipped with xen nowdays? 2014-10-28 12:40:39 could be 2014-10-28 12:40:51 im only checking 64 builder 2014-10-28 12:41:00 will check 32 later 2014-10-28 12:43:32 <[xming]> is the buildbot being forced to rebuild testing/*? 2014-10-28 12:43:33 [xming]: you want minetest to be moved to main? 2014-10-28 12:43:42 <[xming]> yes please 2014-10-28 12:43:57 no not forced 2014-10-28 12:44:21 <[xming]> and I will request seafile + all its deps probably in a few days 2014-10-28 12:44:32 i think when we switched to musl, lots of pkgs got broken in testing. 2014-10-28 12:44:56 <[xming]> as I am going to deploy a cloud for wife and kid later today 2014-10-28 12:44:59 and when we setup the builders, ncopa just enabled continui on fail 2014-10-28 12:45:08 <[xming]> ah 2014-10-28 12:45:30 <[xming]> I only started to have interest in alpine when it switched to musl 2014-10-28 12:45:46 <[xming]> uclibc was grrr 2014-10-28 12:45:53 [xming]: agree 2014-10-28 12:46:39 <[xming]> I told you guys this already, I wanted a musl distro couldn't find one and actually started one, but never finished it 2014-10-28 12:47:11 [xming]: i'm gald you came to alpine 2014-10-28 12:47:14 <[xming]> and a year later I found someone mentioned alpine on /. (I think) in a systemd flame war 2014-10-28 12:47:20 he, bugzilla cant be build. 2014-10-28 12:47:49 <[xming]> it's musl + openrc and very nice people in here 2014-10-28 12:48:16 <[xming]> ncopa: thanks, I am glad that I've found this 2014-10-28 12:49:01 <[xming]> now my new lxc containers are all Alpine, next is the migrating of KVM guests to lxc+alpine 2014-10-28 12:49:25 <[xming]> someday I will probably ditch Gentoo on bare metal and switch to Alpine 2014-10-28 12:49:37 <[xming]> I am just not sure about my desktop 2014-10-28 12:50:09 i love lxc :) 2014-10-28 12:50:16 and my wife of course 2014-10-28 12:50:20 :) 2014-10-28 12:50:50 <[xming]> my wife love lxc, but she doesn't know what it is :p 2014-10-28 12:50:56 :D 2014-10-28 12:51:04 lol 2014-10-28 12:51:07 hmm, could you move some of my packages to main as well? 2014-10-28 12:51:19 <[xming]> since lxc has unpriv. containers, I am all about ditching VMs 2014-10-28 12:51:28 pnutzh4x0r: which package? 2014-10-28 12:51:39 specifically supervisor (and its dependencies), ncmpc, offlineimap, py-tornado (and its dependencies) 2014-10-28 12:52:12 ive been using them for over a month 2014-10-28 12:52:30 whats that builder doing 2014-10-28 12:53:57 ncopa: im also willing to be listed as the maintainer on those packages 2014-10-28 12:55:09 pnutzh4x0r: ok, i'll have a look in a sec 2014-10-28 12:55:11 <[xming]> testing/pptpd and testing/aiccu are working fine here (tm) can be moved to main 2014-10-28 12:57:18 ncopa: http://sprunge.us/iSAQ <- list of packages ive submitted and have been committed 2014-10-28 12:58:05 pnutzh4x0r: and you ahve tested them all? 2014-10-28 12:59:46 yes 2014-10-28 13:00:21 actually, u can drop vimpager 2014-10-28 13:00:31 i dont use that often anymore 2014-10-28 13:00:39 or just keep it in testing 2014-10-28 13:01:44 \o/ 2014-10-28 13:02:15 \o/ 2014-10-28 13:02:33 congrats! 2014-10-28 13:02:39 <[xming]> \o/ 2014-10-28 13:02:41 i dont have x86 builder.. 2014-10-28 13:02:51 <[xming]> ACTION kicks algitbot  2014-10-28 13:03:08 algitbot: please... 2014-10-28 13:03:13 algitbot: please 2014-10-28 13:03:29 <[xming]> clandmeter: lxc? 2014-10-28 13:03:34 looks like it doesnt listen to ops anymore 2014-10-28 13:03:45 <[xming]> bad robot 2014-10-28 13:03:48 yeah i have it 2014-10-28 13:03:58 but not the builder setup 2014-10-28 13:04:09 alpine-edge-x86 2014-10-28 13:04:11 hehe 2014-10-28 13:04:14 i do 2014-10-28 13:05:14 <[xming]> who's the maintainer of the lxc alpine template? someone here or lxc upstream? 2014-10-28 13:05:51 me 2014-10-28 13:05:57 <[xming]> IIRC there was a small fluke, for which I can submit a patch 2014-10-28 13:05:57 i'm the upstream maintainer too i think 2014-10-28 13:06:02 sure 2014-10-28 13:06:14 <[xming]> 'kay 2014-10-28 13:06:28 you can send it to lxc-devel or lxc bugtracker if you want 2014-10-28 13:06:30 and i can ACK it 2014-10-28 13:06:30 <[xming]> ACTION tries to remember what it was 2014-10-28 13:06:38 or show it to me first if you want 2014-10-28 13:06:46 <[xming]> sigh age is my memory's arch enemy 2014-10-28 13:06:52 i know... 2014-10-28 13:07:06 that why i like bugtrackers 2014-10-28 13:07:15 except ours 2014-10-28 13:07:17 :p 2014-10-28 13:07:20 :D 2014-10-28 13:07:35 bugzilla... 2014-10-28 13:07:55 ncopa: does bugzilla have the missing options you are looking for? 2014-10-28 13:08:39 <[xming]> ncopa: /etc/inittab needs this line: `console:12345:respawn:/sbin/getty 38400 console linux` 2014-10-28 13:08:50 <[xming]> otherwise lxc-console doesn't work 2014-10-28 13:09:29 <[xming]> you can just add that, no need to to send to ML and hassles 2014-10-28 13:09:53 lxc-console works for me. 2014-10-28 13:10:16 <[xming]> really, I had to add that line 2014-10-28 13:10:51 <[xming]> must be me then or my gentoo setup 2014-10-28 13:12:27 clandmeter: possible 2014-10-28 13:12:37 i looked at it but i dont htink i was able to figure it out 2014-10-28 13:12:49 but bugzilla is a string candidate indeed 2014-10-28 13:13:14 hm 2014-10-28 13:13:20 it works here too 2014-10-28 13:13:25 i wonder why it happens 2014-10-28 13:13:45 [xming]: what version of lxc is it? 2014-10-28 13:13:46 <[xming]> redmine is nice except it's RoR 2014-10-28 13:13:54 exactly 2014-10-28 13:14:01 thats my exact feeling too 2014-10-28 13:14:11 <[xming]> lxc-1.0.4/ 2014-10-28 13:14:14 hm 2014-10-28 13:14:43 <[xming]> I have used it like 10 years ago for callweaver (OSS pbx, asterisk fork) 2014-10-28 13:15:00 <[xming]> but 768MB wasn't enough for it :( 2014-10-28 13:15:20 <[xming]> these days RAM doesn't matter any more 2014-10-28 13:16:59 <[xming]> ncopa: anything against that I crawl the alpine wiki so it's indexed with yacy? 2014-10-28 13:17:18 yacy? 2014-10-28 13:17:29 feel free to index it 2014-10-28 13:17:33 its good 2014-10-28 13:17:50 <[xming]> yacy is a p2p search engine 2014-10-28 13:17:52 thanks 2014-10-28 13:17:57 yea google told me so 2014-10-28 13:18:02 hmm 2014-10-28 13:18:08 somethign went wrong with git. 2014-10-28 13:18:13 <[xming]> http://yacy.net/en/index.html 2014-10-28 13:18:59 <[xming]> I've already indexed the main alpine page 2014-10-28 13:21:49 clandmeter: whats wrong with git? 2014-10-28 13:22:03 nah, i think i messed up 2014-10-28 13:22:10 wrong commit msg with commit 2014-10-28 13:22:28 ifplugd but it was actually hessleing 2014-10-28 13:22:39 its so nice to see those "uploaded" messages again :) 2014-10-28 13:23:23 with some love, things can be brought back to testing/main. 2014-10-28 13:26:40 <[xming]> oh uploaded, great job guys 2014-10-28 13:29:29 ncopa: have a list from x86? 2014-10-28 13:31:01 http://sprunge.us/VePZ 2014-10-28 13:31:04 i think many are fixed 2014-10-28 13:31:11 its stuck in ocaml atm 2014-10-28 13:31:21 it gives TEXTRELs 2014-10-28 13:31:26 http://bld1.alpinelinux.org/buildlogs/build-edge-x86/testing/ocaml/ocaml-3.12.1-r0.log 2014-10-28 13:32:01 i'm checking what gentoo does 2014-10-28 13:32:01 mine is still building 2014-10-28 13:32:19 they seem to have a patch tarball for ocaml 2014-10-28 13:37:17 could have deleted that one :) 2014-10-28 14:04:54 I just read some backlog... here's a script that checks for bashisms: http://sources.debian.net/src/devscripts/2.14.10/scripts/checkbashisms.pl/ 2014-10-28 14:08:06 [xming]: I run alpine lxc containers version 1.0.6 on an alpine host and lxc-console works without that line in the inittab 2014-10-28 14:08:28 <[xming]> hmm 2014-10-28 14:11:34 [xming]: why aren't you sure if you want to run alpine on your desktop? I use it since 2 months on my notebook (I don't have a desktop anymore) and I'm happy with and it's also my business notebook 2014-10-28 14:11:59 <[xming]> I am so used to my current env 2014-10-28 14:12:27 <[xming]> been a Gentoo users since 2002~2003 before it was even called Gentoo 2014-10-28 14:12:33 <[xming]> anoch or something 2014-10-28 14:12:34 I always have to use lxc-console, because lxc-attach doesn't work because of CAP_SYS_ADMIN :-) 2014-10-28 14:12:44 <[xming]> enouch 2014-10-28 14:12:46 Ah, ok 2014-10-28 14:12:49 <[xming]> meh enoch 2014-10-28 14:12:54 ncopa: dosbox has TEXTREL 2014-10-28 14:13:54 <[xming]> DOS runs in Ring 0 :p 2014-10-28 14:26:23 <[xming]> ncopa: http://wiki.alpinelinux.org/wiki/Seafile:_setting_up_your_own_private_cloud 2014-10-28 14:26:58 ah nice! 2014-10-28 14:39:36 <[xming]> now I lied on the wiki, seafile is not in testing yes :p 2014-10-28 14:40:28 <[xming]> pretty please https://github.com/alpinelinux/aports/pull/5 2014-10-28 14:51:05 i was about to pull it 2014-10-28 14:51:49 im trying to fix ocaml on x86 2014-10-28 14:54:21 [xming]: chown: unknown user/group seafile:seafile 2014-10-28 14:54:43 and you commented the depends. i guess they can go? 2014-10-28 14:55:19 <[xming]> addgroup -S -g 800 seafile 2>/dev/null adduser -h $SEAFILE_HOME -s /bin/sh -G seafile -g seafile -u 800 -D seafile 2>/dev/null 2014-10-28 14:55:27 <[xming]> I have that in the pre-install 2014-10-28 14:55:41 <[xming]> clandmeter: yes they can go, I moved the depends from seafile to seahub 2014-10-28 14:56:00 <[xming]> where are you getting that chown error? 2014-10-28 14:57:34 you didnt add pkgusers/pkggroups 2014-10-28 15:02:52 <[xming]> strange I never had error regarding that (yes I deleted the group and user everytime) 2014-10-28 15:04:57 it doesnt build if i dont add it. 2014-10-28 15:06:52 <[xming]> ah maybe my builder has it, on my test env I deleted for sure 2014-10-28 15:08:16 ncopa: did you see overlayfs is merged? 2014-10-28 15:08:34 into mainline kernel? 2014-10-28 15:08:36 i didnt see 2014-10-28 15:08:37 cool 2014-10-28 15:08:40 \o/ 2014-10-28 15:08:43 finally 2014-10-28 15:08:53 we can make modloop writable on tmpfs :) 2014-10-28 15:09:20 cool 2014-10-28 15:09:25 we can finally consider it yes :) 2014-10-28 15:09:37 we already have it 2014-10-28 15:09:51 just need to modify it for overlayfs 2014-10-28 15:10:10 and enable it in kernel config 2014-10-28 15:10:14 very nice 2014-10-28 15:10:26 yes, forgot about that :) 2014-10-28 15:10:37 jbilyk: hi 2014-10-28 15:10:45 <[xming]> yes in 3.18.rc2 IIR 2014-10-28 15:11:12 hey clandmeter 2014-10-28 15:11:18 anyone heard which will be next longterm release? 2014-10-28 15:11:20 jbilyk: do you know how to fix the broken images on wiki? 2014-10-28 15:11:40 they are broken on chrome based browsers. 2014-10-28 15:11:56 I noticed that recentl 2014-10-28 15:11:58 recently 2014-10-28 15:12:02 no, not sure how to fix it 2014-10-28 15:12:13 some work though 2014-10-28 15:12:42 http://wiki.alpinelinux.org/wiki/1.5x 2014-10-28 15:12:50 thats the url i get :) 2014-10-28 15:13:04 clandmeter: dev tools in chrome on my box says browser is getting 404'd 2014-10-28 15:13:21 i guess its because they are svg 2014-10-28 15:13:32 but one image does work. 2014-10-28 15:13:36 <[xming]> but one of the svg is showing 2014-10-28 15:13:56 ACTION hasn't cared enough to look :) I use the wiki for the text content :P 2014-10-28 15:14:56 ACTION wouldnt care, it wasnt the fp 2014-10-28 15:15:05 clandmeter: maybe we swap out for png? 2014-10-28 15:15:18 but i think they are 2014-10-28 15:15:22 if you look in the src 2014-10-28 15:15:50 http://wiki.alpinelinux.org/w/images/thumb/b/bf/Bookcase.svg/64px-Bookcase.svg.png 2014-10-28 15:16:06 srcset=" 1.5x, /w/images/thumb/b/bf/Bookcase.svg/128px-Bookcase.svg.png 2x" 2014-10-28 15:16:14 ^^ is what's bombing likely 2014-10-28 15:16:34 so instead of scaling the svg properly, chrome is midhandling the tag to think it's loading 1.5 2014-10-28 15:16:40 instead of /w/images/thumb/b/bf/Bookcase.svg/128px-Bookcase.svg.png 2014-10-28 15:17:44 its prefering srcset 2014-10-28 15:17:48 whatever that is 2014-10-28 15:18:20 dunno 2014-10-28 15:18:30 ACTION is a network admin, not web programmer 2014-10-28 15:18:45 you are infra manager ;-) 2014-10-28 15:18:54 or director 2014-10-28 15:19:06 CIO 2014-10-28 15:19:06 <[xming]> where should I use pkgusers/pkgroup? in the APKBUILD? 2014-10-28 15:19:11 def not CIO :P 2014-10-28 15:19:15 that's nangel lol 2014-10-28 15:19:46 [xming]: http://tinyurl.com/ocwkvfk 2014-10-28 15:20:12 i know a bit of html, but that wiki markup hurts my eyes 2014-10-28 15:20:51 <[xming]> so that will make the user/group in the sandbox? 2014-10-28 15:21:07 yes 2014-10-28 15:21:12 <[xming]> with which uid and gid? 2014-10-28 15:21:23 system user 2014-10-28 15:21:29 <[xming]> or not resolved? 2014-10-28 15:21:32 atleast thats what wiki tells me 2014-10-28 15:23:33 clandmeter: http://www.mediawiki.org/wiki/Manual:$wgResponsiveImages 2014-10-28 15:24:05 jbilyk: i saw also some bug about not being able to reset password in wiki 2014-10-28 15:24:42 sounds like this behaviour was introduced in mediawiki 1.21 2014-10-28 15:24:47 sorry it was #3425 2014-10-28 15:25:50 jbilyk: maybe a bug? 2014-10-28 15:26:05 might be - I can try disabling now if nobody objects 2014-10-28 15:26:51 done 2014-10-28 15:26:55 looks like it's fixed 2014-10-28 15:27:09 \o/ 2014-10-28 15:27:12 heh - infra manager actually doing his job for once lol 2014-10-28 15:27:21 <[xming]> clandmeter: ty 2014-10-28 15:27:27 was there a bug report for that one clandmeter ? 2014-10-28 15:27:44 only in my head, but could be. 2014-10-28 15:28:27 searched, can't find one 2014-10-28 15:28:29 <[xming]> thumbs up jbilyk 2014-10-28 15:28:42 <[xming]> now I have pretty icons in chromium 2014-10-28 15:28:46 me too 2014-10-28 15:29:14 everytime i see nilfs-utils i replace the first letter in my mind. 2014-10-28 15:29:19 is it only me? 2014-10-28 15:30:13 yes :) 2014-10-28 15:30:35 :p 2014-10-28 15:33:11 it does build local. 2014-10-28 15:34:55 man you beat me to it... :) 2014-10-28 15:35:05 git reset :p 2014-10-28 15:35:31 git checkout . 2014-10-28 15:36:06 :) ok so you didnt go all the way :) 2014-10-28 15:37:47 <[xming]> with whic lettre? 2014-10-28 15:39:08 [xming]: are you serious? :) 2014-10-28 15:39:42 i htink I'll do dsf2flac x86_64 only 2014-10-28 15:39:45 it wants SSE 2014-10-28 15:39:58 thats ok 2014-10-28 15:40:06 im the only one using it anyway 2014-10-28 15:40:08 i guess 2014-10-28 15:40:13 <[xming]> okay now I am really fustrated, because normally I get this kind of stuffs rather quickly 2014-10-28 15:40:44 the letter preceding the original one. 2014-10-28 15:43:51 <[xming]> ok my 2014-10-28 15:44:16 <[xming]> how come that has never occured to me 2014-10-28 15:44:55 clandmeter: i told you, its just you :-p 2014-10-28 15:45:04 hehe 2014-10-28 15:45:04 <[xming]> they even look alike 2014-10-28 15:45:40 the question would be, how does such utils look like? 2014-10-28 15:47:48 bah 2014-10-28 15:47:54 mongodb has x86_64 disabled 2014-10-28 15:48:00 and now it does not even build with x86 2014-10-28 15:48:03 <[xming]> I meant those 2 lettres 2014-10-28 15:51:39 mongodb is nasty 2014-10-28 15:52:08 \o/ 2014-10-28 15:52:09 \o/ 2014-10-28 15:52:12 finally 2014-10-28 15:52:19 succesfull day 2014-10-28 15:52:23 yes! 2014-10-28 15:52:30 got some digital eyes 2014-10-28 15:52:57 https://mjanja.ch/wordpress/wp-content/uploads/2013/09/borat_great_success.jpg 2014-10-28 15:53:00 ACTION claps ncopa and clandmeter  2014-10-28 15:53:56 [xming]: did you submit chromium already? 2014-10-28 15:56:53 now if there was a DE with some more eyecandy, i would switch my destkop to alpine. 2014-10-28 15:57:12 ncopa: dont forget about those lua aports 2014-10-28 15:58:43 which ones? 2014-10-28 15:58:51 the one with lua in depends 2014-10-28 15:59:07 i fix building them by adding it to makedeps 2014-10-28 15:59:22 but i guess they will have issues when you run them. 2014-10-28 16:01:25 pnutzh4x0r: could you please check what license supervisor is? 2014-10-28 16:04:46 ok time to go 2014-10-28 16:04:54 i will set up the v3.1 builders tomorrow 2014-10-28 16:14:17 [xming]: sorry i thouht you were the one packaging chromium... 2014-10-28 16:14:27 nick concliction 2014-10-28 16:15:40 s/c/f 2014-10-28 16:15:52 its time to go home it seems. 2014-10-28 18:12:38 ncopa: https://github.com/Supervisor/supervisor/blob/master/LICENSES.txt 2014-10-28 18:12:54 ncopa: looks like supervisor is a mish-mash of BSD-like licences 2014-10-28 18:13:15 ncopa: https://pypi.python.org/pypi/supervisor lists it as BSD-derived 2014-10-28 18:34:12 do you happen to know what kind of black magic I need to perform to bridge wireless network to LXC on my laptop? 2014-10-28 18:34:28 clandmeter, ncopa ↑ 2014-10-28 18:34:46 barthalion: bridging the veth and the wlan interface doesn't work? 2014-10-28 18:35:12 i guess it works the same as a normal interface? 2014-10-28 18:35:25 I'm rather clueless because I completely suck at networking, so whatever you say to me, I'll try it until the week ends 2014-10-28 18:35:27 i know from openwrt routers that the wifi driver has to support bridging 2014-10-28 18:35:50 oh, that might be it 2014-10-28 18:36:12 i never tried it, so it could be. 2014-10-28 18:36:21 barthalion: did you folow http://wiki.alpinelinux.org/wiki/LXC ? 2014-10-28 18:36:30 barthalion: if bridging doesn't work you can route the traffic 2014-10-28 18:36:55 clandmeter: I'm not on alpine, but I'll take a look, ty 2014-10-28 18:36:59 jomat: noted, thanls 2014-10-28 18:37:01 thanks even 2014-10-28 18:37:08 if you still need the same network address in the containers as your wifi, you have to enable arp proxying 2014-10-28 18:37:56 barthalion: you need to pay for non alpine related questions. 2014-10-28 18:38:00 :) 2014-10-28 18:38:08 oh boy, I wonder if it wouldn't be faster to buy a longer cable 2014-10-28 18:38:25 i will deduct it from your alpine salary 2014-10-28 18:38:44 clandmeter: you can't pay me less than you do now! 2014-10-28 18:41:22 barthalion: true, we still have to allocate the budget. 2014-10-28 20:17:13 <[xming]> barthalion: you need a wireless NIC which can do monitor mode 2014-10-28 20:18:10 <[xming]> the old orinoco gold cards could do that 2014-10-28 20:18:34 <[xming]> otherwise some brc, don't know about alteros 2014-10-28 20:18:56 <[xming]> but openwrt has something called relayd to do that for all cards 2014-10-28 20:19:19 <[xming]> haven't looked into how they do that, I suspect proxy arp 2014-10-28 21:16:59 <[xming]> http://nbd.name/gitweb.cgi?p=relayd.git;a=summarybarthalion: 2014-10-28 21:17:10 <[xming]> barthalion: ^^^ 2014-10-28 21:17:27 <[xming]> it's indeed proxy arping 2014-10-28 21:23:31 [xming]: thanks 2014-10-28 21:30:53 <[xming]> speaking of proxy arp, I once encountered a Cisco router configured by probably someone who never heard about default gateway, so it was configured to proxy arp the whole internet 2014-10-28 21:31:32 <[xming]> obviously once the table is full they had tons of strange issues, it was fun 2014-10-28 21:40:22 <[xming]> barthalion: list of wifi nic supporting monitor mode http://www.aircrack-ng.org/doku.php?id=compatibility_drivers 2014-10-28 23:52:08 hello again alpine-devel 2014-10-28 23:53:38 i managed to compile a new subpackage of collectd with minimal modifications to the APKBUILD file (although i did need a lot abuild -r , abuild -R and apk add before it compiled) 2014-10-28 23:53:49 i was wondering how to submit this ? 2014-10-28 23:54:19 oh, i did that in 2.7 but it should not male a difference 2014-10-28 23:54:22 make 2014-10-29 00:08:21 pierre76: u can send a git diff to the mailing list 2014-10-29 00:08:31 or u can do a pull request on github 2014-10-29 00:08:56 alternatively, u can sprunge the diff and leave it for ncopa or some of the other maintainers 2014-10-29 00:19:43 git diff does not seem very human friendly on my screen.. 2014-10-29 00:20:02 sprunge ? 2014-10-29 00:22:57 ok git diff looks better in a file, i'll send that to alpine-devel if i can without being subscribed ? 2014-10-29 00:33:33 done, thanks for the help 2014-10-29 06:24:16 why this guy is alwais coming when almost everyone sleeps 2014-10-29 08:08:44 <[xming]> BST != +1 http://wiki.alpinelinux.org/wiki/Alpine_Linux:IRC_Availability 2014-10-29 08:08:52 <[xming]> BST+DST == +1 2014-10-29 08:49:16 BST = British Summer Time ? (+1) ? 2014-10-29 08:50:12 and BST = Bangladesh Standard Time ( +6) 2014-10-29 08:54:27 hmm so CET + DaylightSaving = ? 2014-10-29 08:58:03 coz DST not always = 1 ? 2014-10-29 09:02:10 <[xming]> BST is British standard time, which is GMT/UTC during winter 2014-10-29 09:02:32 <[xming]> BST+DST == +1 2014-10-29 09:02:44 CET is always +1 2014-10-29 09:02:46 <[xming]> CET == +2 2014-10-29 09:02:48 <[xming]> CET == +1 2014-10-29 09:02:54 <[xming]> CET+DST == +2 2014-10-29 09:03:01 but now we have daylight saving time, so it's +2 2014-10-29 09:03:10 <[xming]> it's just over 2014-10-29 09:03:14 <[xming]> it's +1 now :p 2014-10-29 09:03:29 ah, right 2014-10-29 09:03:30 <[xming]> at least in this part of CET 2014-10-29 09:03:32 gimme more coffee 2014-10-29 09:04:01 <[xming]> vkrishn: DST is Day light Saving time 2014-10-29 09:04:22 <[xming]> DST is the most part of the world is ++1 2014-10-29 09:04:31 <[xming]> at least in EU 2014-10-29 09:07:28 ok, removed BST confusion 2014-10-29 09:10:10 may add a line that DST calc not done 2014-10-29 09:10:57 I thought BST = GMT/UTC + DST 2014-10-29 09:11:06 anyway 2014-10-29 09:12:11 rm -rf DST 2014-10-29 09:12:32 barthalion, I learned about DST from a polish friend ;) 2014-10-29 09:12:50 more coffee to you 2014-10-29 09:14:55 yeah, there has been lots news about scrapping DST 2014-10-29 09:16:09 how does it save it is supposedly energy efficient world ? 2014-10-29 09:16:59 with LED 10w = 60Watt bulb ;-)) 2014-10-29 09:17:18 ever been to africa? 2014-10-29 09:17:42 nope 2014-10-29 09:17:58 10w still too much for their batteries :) 2014-10-29 09:18:17 but why batteries, use solar 2014-10-29 09:18:29 two 3w panels 2014-10-29 09:18:34 or even 1 2014-10-29 09:19:13 there has been lot of NGO program/initiative that distributes such lamps 2014-10-29 09:19:36 why use led lamp where there is sun? 2014-10-29 09:20:01 it charges in day time ;)))) 2014-10-29 09:20:17 of course battery 2014-10-29 09:21:07 but closed indoors could be direct 2014-10-29 09:21:19 like huts 2014-10-29 09:23:07 maybe use nice capacitors 2014-10-29 09:38:53 <[xming]> vkrishn: DST is not a timezone perse 2014-10-29 09:39:34 <[xming]> BST is just BST which is GMT/UTC when DST is not applied 2014-10-29 09:39:39 yes, its +1 (mostly) to whatever timezone 2014-10-29 09:39:53 <[xming]> yeah 2014-10-29 09:40:00 so UTC+DST = BST 2014-10-29 09:40:12 <[xming]> vkrishn: no :( 2014-10-29 09:40:28 <[xming]> right now BST == GMT == UTC 2014-10-29 09:40:30 oh, just funny BST 2014-10-29 09:40:37 nice 2014-10-29 09:40:39 thanks 2014-10-29 09:40:56 <[xming]> during summer BST + DST == GMT +1 == UTC +1 2014-10-29 09:41:52 so why add word "summer" to it ahhh 2014-10-29 09:41:55 <[xming]> vkrishn: you must've talked to some Brits during the summer, to get the impression BST == UTC +1 :p 2014-10-29 09:42:21 <[xming]> becuase DST is only applies during "summer" 2014-10-29 09:42:37 <[xming]> between end of March till end of October 2014-10-29 09:42:46 and still call BST == UTC after DST is over 2014-10-29 09:42:53 <[xming]> yes 2014-10-29 09:43:02 <[xming]> actaqully BST is always UTC 2014-10-29 09:43:10 so in winter its BST (0) 2014-10-29 09:43:13 <[xming]> BST + DST == UTC + 1 2014-10-29 09:43:15 ;) 2014-10-29 09:43:36 <[xming]> vkrishn: you are getting it :D 2014-10-29 09:43:49 gee, really need to scrape DST 2014-10-29 09:43:54 <[xming]> but most people omit DST during summer 2014-10-29 09:44:02 <[xming]> vkrishn: are you from the States? 2014-10-29 09:44:10 india 2014-10-29 09:44:16 <[xming]> okay 2014-10-29 09:44:24 <[xming]> so you have no notion of DST? 2014-10-29 09:45:12 <[xming]> it's really confusing if you never grow up in place with DST 2014-10-29 09:46:37 wonder how they program their automated coffee machine 2014-10-29 09:46:44 in morning 2014-10-29 09:47:13 beep, beep, coffee at 7am or 8am siiir 2014-10-29 09:55:30 ncopa, have a look at https://github.com/01org/dpdk-ovs 2014-10-29 09:55:35 Open vSwitch accelerated by Intel-DPDK 2014-10-29 09:57:27 vkrishn: after v3.1 2014-10-29 09:57:33 i want wrap up v3.1 now 2014-10-29 09:57:37 not add new features 2014-10-29 09:57:50 ok 2014-10-29 09:58:29 which kernel will it have? 2014-10-29 11:09:13 ncopa: how about moving syslog-ng to main now? 2014-10-29 12:05:42 ncopa: im seeing this in xorg log 2014-10-29 12:05:44 [ 121.452] (EE) Failed to load /usr/lib/xorg/modules/drivers/modesetting_drv.so: Error relocating /usr/lib/xorg/modules/drivers/modesetting_drv.so: ms_shadowUpdate32to24: symbol not found 2014-10-29 12:15:10 i think that is what dalias had 2014-10-29 12:18:09 im trying to start slim 2014-10-29 12:18:22 at home it was working, but here on vmware, it keeps restarting 2014-10-29 12:18:35 wierd 2014-10-29 12:18:41 weird* 2014-10-29 12:19:00 ah 2014-10-29 12:19:02 not weird 2014-10-29 12:19:16 ? 2014-10-29 12:19:23 i know whats wrong 2014-10-29 12:19:43 its the patch i added from fedora 2014-10-29 12:20:18 http://git.alpinelinux.org/cgit/aports/tree/main/xf86-video-modesetting/0001-modesetting-24bpp-are-too-confusing-shadow-our-way-o.patch 2014-10-29 12:20:22 the problem is 2014-10-29 12:20:37 http://git.alpinelinux.org/cgit/aports/tree/main/xf86-video-modesetting/0001-modesetting-24bpp-are-too-confusing-shadow-our-way-o.patch#n371 2014-10-29 12:20:54 the missing symbol is added with a separate .c file 2014-10-29 12:21:32 the file is added to Makefile.am to get built 2014-10-29 12:21:35 http://git.alpinelinux.org/cgit/aports/tree/main/xf86-video-modesetting/0001-modesetting-24bpp-are-too-confusing-shadow-our-way-o.patch#n33 2014-10-29 12:21:36 but 2014-10-29 12:21:44 Makefile.in is not regenerated 2014-10-29 12:21:58 so it never link in sh3224.o 2014-10-29 12:24:02 not sure the slim issue is related to modesetting. 2014-10-29 12:24:25 (EE) Failed to load /usr/lib/xorg/modules/drivers/modesetting_drv.so: 2014-10-29 12:24:33 that means it fails to load the driver 2014-10-29 12:24:36 likely related 2014-10-29 12:24:39 i'll fix it 2014-10-29 12:30:02 clandmeter: can you please check if it solves your issue? 2014-10-29 12:30:36 yes when its ready building 2014-10-29 12:35:17 looks like it having problems downloading cairo 2014-10-29 12:36:18 how come x86 log files open in browser, but 64 doesnt. 2014-10-29 12:36:45 open in browser? 2014-10-29 12:36:56 i can read it in the browser 2014-10-29 12:36:59 i know why.. 2014-10-29 12:37:04 # remove line below to disable colors 2014-10-29 12:37:04 USE_COLORS=1 2014-10-29 12:37:06 escape chars 2014-10-29 12:37:12 can rsync enabled to logs ? 2014-10-29 12:37:53 ah ok. can you disable colors? 2014-10-29 12:37:53 vkrishn: currently no 2014-10-29 12:37:59 I can pull/sync only onces that I want, or just check if they changed 2014-10-29 12:37:59 clandmeter: i just did 2014-10-29 12:38:02 good catch 2014-10-29 12:38:04 thx 2014-10-29 12:38:07 ok 2014-10-29 12:38:47 finanly dl done. 2014-10-29 12:38:53 now 10minutes of configure 2014-10-29 12:41:03 ncopa: modesetting ok 2014-10-29 12:41:09 slim: not ok 2014-10-29 12:44:03 lxdm is the same 2014-10-29 12:46:08 ncopa: seen this before? http://sprunge.us/jhij 2014-10-29 12:46:40 grsec kernel? 2014-10-29 12:46:55 no 2014-10-29 12:47:03 grsec doesnt boot 2014-10-29 12:47:25 1G memory should be enough right? 2014-10-29 12:48:10 yes 2014-10-29 12:48:13 hm dunno 2014-10-29 12:53:37 really weird, it was working not so long ago. 2014-10-29 13:06:16 I think logs in build-edge-x86_64 may be having escape chars 2014-10-29 13:14:19 vkrishn: i think i just fixed it 2014-10-29 13:15:38 ok 2014-10-29 13:29:50 ncopa: on 3.0 it still works. so something must have changed in edge to make it not functional on vmware. 2014-10-30 08:52:19 update rsyslog (8.4.2), monit (5.10) pls 2014-10-30 08:56:53 ncopa: i just read other channel about dbg pkg... 2014-10-30 08:57:14 next version of nanomsg may possibly have websocket 2014-10-30 10:41:20 i am bootstrapping the build v3.1 build servers 2014-10-30 13:05:20 ncopa: you probably missed my question yesterday: how about syslog-ng and move it to main? any complains left? 2014-10-30 13:10:30 i saw it, but forgot while beeing busy with something else, sorry 2014-10-30 13:40:48 i c 2014-10-30 13:40:50 no problem 2014-10-30 15:29:57 ncopa: 3.1 coming along? 2014-10-30 15:32:07 yup 2014-10-30 15:34:02 lots of breakage? 2014-10-30 19:13:35 clandmeter: hi, chromium APKBUILD -> http://sprunge.us/CWOD 2014-10-30 19:13:51 patch -> http://sprunge.us/FjWH 2014-10-30 19:14:06 but not ready yet, special for you :) 2014-10-30 19:14:19 i can run it only without sandboc 2014-10-30 19:14:24 *sandbox 2014-10-30 19:14:35 with --no-sandbox option 2014-10-30 19:15:46 and also should get our own google api keys... 2014-10-30 19:17:03 alos should build with latest libwebp-dev package (because demux.h) 2014-10-31 10:21:28 <[xming]> seafile + seahub and their dependencies are working fine here, maybe time to move them to main? 2014-10-31 10:21:56 <[xming]> do I just make a PR for that? 2014-10-31 10:24:25 [xming]: if you specifiy the deps, we can just move them. 2014-10-31 10:24:38 thats probably easier then creating a pr. 2014-10-31 10:28:31 <[xming]> http://sprunge.us/YOID 2014-10-31 10:30:09 <[xming]> http://sprunge.us/IIJM 2014-10-31 10:54:35 [xming]: i think you missed a few deps :) 2014-10-31 11:01:49 <[xming]> clandmeter: silly me, I just pasted my aports, there are of course others which are also in testing 2014-10-31 11:18:28 [xming]: no problem. it should be fixed now. 2014-10-31 12:37:51 <[xming]> clandmeter: ty 2014-10-31 14:03:48 ncopa, I am seeing mqtt topic like, 2014-10-31 14:03:50 git/acf-freeradius3/0.0-stable 2014-10-31 14:03:59 and, git/acf-freeradius3/0.0-stable/last5 2014-10-31 14:04:16 hope its ok ? 2014-10-31 14:08:20 00-stable? 2014-10-31 14:09:59 i suppose acf-freeradisu3 has a git branch called 0.0-stable then 2014-10-31 14:10:38 I mean considering the last part of topic to be either category OR message name 2014-10-31 14:11:29 it just a practice 2014-10-31 14:11:32 it is 2014-10-31 14:11:51 and dont't think mqtt community has any thought on it 2014-10-31 14:12:27 so, git/acf-freeradius3/0.0-stable would be git/acf-freeradius3/0.0-stable/current 2014-10-31 14:13:01 and git/acf-freeradius3/0.0-stable would be considered as category 2014-10-31 14:13:34 kinda filesystem 2014-10-31 14:13:50 but then no norm on it yet 2014-10-31 14:15:06 did nl.alpinelinux.org fall down? 2014-10-31 14:16:54 I can ping it though 2014-10-31 14:17:06 nope 2014-10-31 14:17:23 * clandmeter has quit (Ping timeout: 245 seconds) 2014-10-31 14:28:45 online now 2014-10-31 14:29:27 for a moment I thought , halloween break 2014-10-31 14:33:10 looks like some hickup on our provider side 2014-10-31 14:33:24 seems nl was partially down 2014-10-31 14:36:16 we noticed nl.a.o was down 2014-10-31 14:36:41 yeah i couldnt even go to google.com on my phone... 2014-10-31 15:27:01 vkrishn / ncopa: yes, 0.0-stable is a strange branch name for acf-freeradius3 2014-10-31 15:27:26 but, that's what happened when I was asked to backport a new package that needed changes to run on alpine 2.7 :/ 2014-10-31 15:27:49 it was originally developed for alpine 3 2014-10-31 18:51:05 clandmeter: ping 2014-10-31 21:06:49 k0r10n: pong